Generalized Anxiety Disorder: Advances in Research and Practice

GENERALIZED ANXIETY DISORDER

Generalized Anxiety Disorder ADVANCES IN RESEARCH AND PRACTICE

Edited by Richard G. Heimberg Cynthia L. Turk Douglas S. Mennin

THE GUILFORD PRESS New York London

© 2004 The Guilford Press A Division of Guilford Publications, Inc. 72 Spring Street, New York, NY 10012 All rights reserved No part of this book may be reproduced, translated, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, microfilming, recording, or otherwise, without written permission from the Publisher. Printed in the United States of America This book is printed on acid-free paper. Last digit is print number: 9 8 7 6 5 4 3 2 1 Library of Congress Cataloging-in-Publication Data Generalized anxiety disorder : advances in research and practice / edited by Richard G. Heimberg, Cynthia L. Turk, Douglas S. Mennin. p. cm. Includes bibliographical references and index. ISBN 1-57230-972-5 1. Anxiety. I. Heimberg, Richard G. II. Turk, Cynthia L. III. Mennin, Douglas S. RC531.G465 2004 616.85¢22—dc22 2003015738

About the Editors

About the Editors

Richard G. Heimberg, PhD, is Professor of Psychology, Director of Clinical Training, and Director of the Adult Anxiety Clinic at Temple University. He is well known for his efforts in developing and evaluating cognitive-behavioral treatments, especially for social anxiety. Dr. Heimberg has published more than 225 articles and chapters on anxiety disorders and related topics. He is coeditor or coauthor of several books, including Social Phobia: Diagnosis, Assessment, and Treatment (1995, Guilford Press) and Cognitive-Behavioral Group Therapy for Social Phobia: Basic Mechanisms and Clinical Strategies (2002, Guilford Press). Dr. Heimberg is past president of the Association for Advancement of Behavior Therapy. He was recently named one of the four most inf luential psychological researchers in anxiety in a survey of members of the Anxiety Disorders Association of America, and he serves on the Scientific Advisory Board of that association. Dr. Heimberg is a founding fellow of the Academy of Cognitive Therapy and was the inaugural recipient of the Academy’s A. T. Beck Award for Significant and Enduring Contribution to Cognitive Therapy. Cynthia L. Turk, PhD, joined the psychology faculty at La Salle University in 2002 and is currently the Director of the La Salle Anxiety Disorders Treatment and Research Program. She received her doctoral degree from Oklahoma State University in 1996. Prior to her current position, Dr. Turk was Associate Director of the Adult Anxiety Clinic at Temple University and Director of its Generalized Anxiety Program. Her main interests are in the assessment and treatment of anxiety disorders, particularly generalized anxiety disorder and social phobia. Douglas S. Mennin, PhD, is currently Assistant Professor in the Department of Psychology at Yale University and Director of Yale Anxiety and Mood Services. He received his doctoral degree from Temple University in 2001, and with Drs. Heimberg and Turk, helped to establish the Generalized Anxiety Program at the Adult Anxiety Clinic at Temple University. Dr. Mennin’s primary interests are generalized anxiety disorder, the role of difficulties in emotion regulation in generalized anxiety disorder and other mental disorders, and the integration of cognitivebehavioral and emotion-focused therapies. v

Contributors

Contributors

Anne Marie Albano, PhD, Child Study Center, New York University School of Medicine, New York, New York Oscar M. Alcaine, MA, Department of Psychology, Pennsylvania State University, University Park, Pennsylvania Patricia M. Averill, PhD, University of Texas–Houston Harris County Psychiatric Center, Houston, Texas J. Gayle Beck, PhD, Department of Psychology, University at Buffalo, State University of New York, Buffalo, New York Evelyn Behar, MS, Department of Psychology, Pennsylvania State University, University Park, Pennsylvania Thomas D. Borkovec, PhD, Department of Psychology, Pennsylvania State University, University Park, Pennsylvania Kristin Buhr, MA, Department of Psychology, Concordia University, Montreal, Quebec, Canada Louis G. Castonguay, PhD, Department of Psychology, Pennsylvania State University, University Park, Pennsylvania Katherine Crits-Christoph, PhD, Center for Psychotherapy Research, Department of Psychiatry, University of Pennsylvania, Philadelphia, Pennsylvania Paul Crits-Christoph, PhD, Center for Psychotherapy Research, Department of Psychiatry, University of Pennsylvania, Philadelphia, Pennsylvania Michel J. Dugas, PhD, Department of Psychology, Concordia University, Montreal, Quebec, Canada vii

viii

Contributors

Mary Beth Connolly Gibbons, PhD, Center for Psychotherapy Research, Department of Psychiatry, University of Pennsylvania, Philadelphia, Pennsylvania Robert A. Gould, PhD, (deceased) Department of Psychiatry, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts Jack M. Gorman, MD, Department of Psychiatry, Mt. Sinai Medical School, New York, New York Sabine Hack, MD, Child Study Center, New York University School of Medicine, New York, New York Richard G. Heimberg, PhD, Adult Anxiety Clinic, Department of Psychology, Temple University, Philadelphia, Pennsylvania Jennifer L. Hudson, PhD, Department of Psychology, Macquarie University, Sydney, Australia Ronald C. Kessler, PhD, Department of Health Care Policy, Harvard Medical School, Boston, Massachusetts Robert Ladouceur, PhD, Ecole de Psychologie, Université Laval, Ste-Foy, Quebec, Canada Robert L. Leahy, PhD, American Institute for Cognitive Therapy, New York, New York R. Bruce Lydiard, MD, Department of Neuropsychiatry and Behavioral Science, University of South Carolina School of Medicine, Columbia, South Carolina, and Southeast Health Consultants, Charleston, South Carolina Colin MacLeod, PhD, Department of Psychology, University of Western Australia, Nedlands, Perth, Australia Douglas S. Mennin, PhD, Department of Psychology, Yale University, New Haven, Connecticut Jan Mohlman, PhD, Department of Psychology, Syracuse University, Syracuse, New York Christine Molnar, PhD, Brain Stimulation Laboratory, Medical University of South Carolina, Charleston, South Carolina Jeannine Monnier, PhD, Department of Psychiatry and Behavioral Sciences, Medical University of South Carolina, Charleston, South Carolina Michelle G. Newman, PhD, Department of Psychology, Pennsylvania State University, University Park, Pennsylvania

Contributors

ix

Michael W. Otto, PhD, Department of Psychiatry, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts Ronald M. Rapee, PhD, Department of Psychology, Macquarie University, Sydney, Australia Elizabeth Rutherford, PhD, Department of Psychology, University of Western Australia, Nedlands, Perth, Australia Steven A. Safren, PhD, Department of Psychiatry, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts Smit S. Sinha, MD, Department of Psychiatry, Columbia University, New York, New York Cynthia L. Turk, PhD, Department of Psychology, La Salle University, Philadelphia, Pennsylvania Ellen E. Walters, MS, Department of Health Care Policy, Harvard Medical School, Boston, Maasachusetts David O’Neill Washington, PhD, Department of Psychiatry, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts Adrian Wells, MSc, PhD, Department of Clinical Psychology, University of Manchester, Manchester, United Kingdom Hans-Ulrich Wittchen, PhD, Institute of Clinical Psychology and Psychotherapy, Technical University of Dresden, Dresden, Germany, and Max Planck Institute of Psychiatry, Clinical Psychology and Epidemiology, Munich, Germany

Preface

Preface

One might say that we are living in an “age of anxiety” (Twenge, 2000, p. 1007). Given an unstable economy, war, terrorist threats, shooting sprees, and the renewed spread of deadly disease, it may seem that our sense of security has diminished. Indeed, it appears that anxiety may be a more prominent reaction in modern times: Anxiety levels have increased steadily in birth cohorts over the past 50 years (Twenge, 2000). With this rise in anxiety has come an increase in our understanding of its adaptive and maladaptive aspects. Most notably, the study of the pathological forms of anxiety—anxiety disorders—has increased tremendously over the past 20 years (Cox, Wessel, Norton, Swinson, & Direnfeld, 1995; Norton, Cox, Asmundson, & Maser, 1995). Anxiety disorders have been shown to be associated with functional disability, poor quality of life, and increased health care costs, and are experienced in some form by over 20% of the population (Kessler et al., 1994). These findings suggest a significant public health care concern. In response to the clinical importance of anxiety, theories about pathological mechanisms have been generated and have demonstrated considerable empirical support (e.g., Barlow, 2002; Clark, 1986; Rapee & Heimberg, 1997). Novel treatments also have been developed that specifically address these difficulties and have been shown to be quite effective in treating anxiety disorders and ameliorating the suffering of those aff licted with them (Barlow, Gorman, Shear, & Woods, 2000). One notable exception to this growth is the study and treatment of generalized anxiety disorder (GAD). Unlike panic disorder (e.g., McNally, 1994), social phobia (e.g., Heimberg & Becker, 2002; Heimberg, Liebowitz, Hope, & Schneier, 1995), or obsessive–compulsive disorder (Swinson, Antony, Rachman, & Richter, 1998), in which considerable advances are clearly evident, GAD remains an understudied (Dugas, xi

xii

Preface

2000), misunderstood (Persons, Mennin, & Tucker, 2001), and treatmentresistant (Borkovec & Ruscio, 2001) disorder. Much of the confusion surrounding the disorder comes from its checkered past as a residual diagnosis in the third edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-III; American Psychiatric Association, 1980). This early form of the diagnosis resulted in poor reliability and led many to question its validity. In recent years, the reliability of GAD has improved considerably (Brown, DiNardo, Lehman, & Campbell, 2001; Starcevic & Bogojevic, 1999). With this improvement and better definition of the key features of GAD, research into its nature and treatment has begun to catch up to research on the other anxiety disorders. As will be demonstrated throughout this book, GAD is now recognized as a widely prevalent disorder associated with substantial functional impairment and reduced life satisfaction. Thus it is appropriate that the field has turned its attention to this important problem. We are excited to have the opportunity to present this volume, one of the first of its kind to focus on GAD. We have brought together many of the world’s leading authorities to address the phenomenology, etiology, pathological mechanisms, and treatment of GAD. These investigators and clinicians shed light on this important but neglected syndrome and help provide direction for future inquiry. In the book’s chapters, the contributors present state-of-the-art reviews and analyses of almost every aspect of GAD. We believe that this work will prove useful to both novices and experts and to both researchers and clinicians interested in GAD. The book is organized into four sections. In Part I (“Generalized Anxiety Disorder in the Clinic and in the Community”), the three of us examine the history of the diagnosis of GAD and its essential feature, excessive and uncontrollable worry. Kessler, Walters, and Wittchen provide an in-depth analysis of the epidemiology of the disorder. Hudson and Rapee examine the role of temperament and environment in the development of GAD. In Part II (“Theoretical and Empirical Approaches”), five chapters examine diverse perspectives on the understanding of GAD. Borkovec, Alcaine, and Behar present the most comprehensive review to date on the role of worry in GAD. MacLeod and Rutherford provide a similarly detailed analysis of the information-processing problems that are evident in persons with GAD. Dugas, Buhr, and Ladouceur examine the role of intolerance of uncertainty in the etiology and maintenance of GAD, and Wells gives us a comprehensive account of his cognitive model of GAD. From a very different, and equally important, perspective, Sinha, Mohlman, and Gorman provide an up-to-the-minute examination of the rapidly developing study of the neurobiology of GAD. Part III (“Assessment and Treatment”) begins with a detailed presentation of the assessment of worry and GAD, again by the three of us.

Preface

xiii

Gould and colleagues present a meta-analytic review of various forms of cognitive-behavioral treatment (CBT) for GAD, and Leahy provides a valuable perspective on the application of one form of CBT to patients with GAD in clinical practice. Crits-Christoph and colleagues describe their supportive–expressive psychodynamic therapy for GAD and present some new data on its efficacy. Newman and colleagues provide a detailed description of their integrative approach to treatment of GAD. Lydiard and Monnier do the same for pharmacotherapy of GAD. Part IV is devoted to “Special Populations.” Albano and Hack give us a complete and thorough look at the assessment and treatment of GAD in children and adolescents. Beck and Averill do the same for GAD in older adults. These are areas that call for our greatest attention. Generalized Anxiety Disorder: Advances in Research and Practice is a comprehensive guide to the understanding and treatment of GAD. However, there is still much to learn, so research continues. We hope that this volume will stimulate you to think about GAD in new ways, and lead to new and productive directions in treatment and research on this important disorder. Many people have played an important part in the development of this book, and we can never adequately express our gratitude to all of them. Thanks to the chapter contributors. Thanks to Barbara Watkins of The Guilford Press, who helped us in many ways from the first thought to the final product. Thanks to David M. Fresco for his many contributions to our thinking about GAD. Thanks to Ismael Alvarez, administrative assistant to Richard G. Heimberg at the Adult Anxiety Clinic of Temple University, for his many contributions to this book. Most of all, thanks to the patients with GAD who have let us learn from them as we have attempted to help them. RICHARD G. HEIMBERG CYNTHIA L. TURK DOUGLAS S. MENNIN

REFERENCES American Psychiatric Association. (1980). Diagnostic and statistical manual of mental disorders (3rd ed.). Washington, DC: Author. Barlow, D. H. (2002). Anxiety and its disorders (2nd ed.). New York: Guilford Press. Barlow, D. H., Gorman, J. M., Shear, K. M., & Woods, S. W. (2000). Cognitive-behavioral therapy, imipramine, or their combination for panic disorder: A randomized controlled trial. Journal of the American Medical Association, 283, 2529–2536. Borkovec, T. D., & Ruscio, A. M. (2001). Psychotherapy for generalized anxiety disorder. Journal of Clinical Psychiatry, 62, 37–42. Brown, T. A., DiNardo, P. A., Lehman, C. L., & Campbell, L. A. (2001). Reliability of

xiv

Preface

DSM-IV anxiety and mood disorders: Implications for the classification of emotional disorders. Journal of Abnormal Psychology, 110, 49–58. Clark, D. M. (1986). A cognitive approach to panic. Behaviour Reserach and Therapy, 24, 461–470. Cox, B. J., Wessel, I., Norton, G. R., Swinson, R. P., & Direnfeld, D. M. (1995). Publication trends in anxiety disorders research: 1990–1992. Journal of Anxiety Disorders, 9, 531–538. Dugas, M. J. (2000). Generalized anxiety disorder publications: So where do we stand? Journal of Anxiety Disorders, 14, 31–40. Heimberg, R. G., & Becker, R. E. (2002). Cognitive-behavioral group therapy for social phobia: Basic mechanisms and clinical strategies. New York: Guilford Press. Heimberg, R. G., Liebowitz, M. R., Hope, D. A., & Schneier, F. R. (Eds.). (1995). Social phobia: Diagnosis, assessment, and treatment. New York: Guilford Press. Kessler, R. C., McGonagle, K. A., Zhao, S., Nelson, C. B., Hughes, M., Eshleman, S., Wittchen, H. U., & Kendler, K. S. (1994). Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in the United States: Results from the National Comorbidity Survey. Archives of General Psychiatry, 51, 8–19. McNally, R. J. (1994). Panic disorder: A critical analysis. New York: Guilford Press. Norton, G. R., Cox, B. J., Asmundson, G. J. G., & Maser, J. D. (1995). The growth of research on anxiety disorders in the 1980s. Journal of Anxiety Disorders, 9, 75–85. Persons, J. B., Mennin, D. S., & Tucker, D. E. (2001). Common misconceptions about the nature and treatment of generalized anxiety disorder. Psychiatric Annals, 31, 501–508. Rapee, R. M., & Heimberg, R. G. (1997). A cognitive-behavioural model of anxiety in social phobia. Behaviour Research and Therapy, 35, 741-756. Starcevic, V., & Bogojevic, G. (1999). The concept of generalized anxiety disorder: Between the too narrow and too wide diagnostic criteria. Psychopathology, 32, 5–11. Swinson, R. P., Antony, M. M., Rachman, S., & Richter, M. A. (Eds.). (1998). Obsessive– compulsive disorder: Theory, research, and treatment. New York: Guilford Press. Twenge, J. M. (2000). The age of anxiety?: Birth cohort change in anxiety and neuroticism 1952–1993. Journal of Personality and Social Psychology, 79, 1007–1021.

Contents

Contents

PART I. Generalized Anxiety Disorder in the Clinic and in the Community 1. Clinical Presentation and Diagnostic Features Douglas S. Mennin, Richard G. Heimberg, and Cynthia L. Turk

3

2. Epidemiology Ronald C. Kessler, Ellen E. Walters, and Hans-Ulrich Wittchen

29

3. From Anxious Temperament to Disorder: An Etiological Model Jennifer L. Hudson and Ronald M. Rapee

51

PART II. Theoretical and Empirical Approaches 4. Avoidance Theory of Worry and Generalized Anxiety Disorder Thomas D. Borkovec, Oscar M. Alcaine, and Evelyn Behar

77

5. Information-Processing Approaches: Assessing the Selective Functioning of Attention, Interpretation, and Retrieval Colin MacLeod and Elizabeth Rutherford

109

6. The Role of Intolerance of Uncertainty in Etiology and Maintenance Michel J. Dugas, Kristin Buhr, and Robert Ladouceur

143

7. A Cognitive Model of GAD: Metacognitions and Pathological Worry Adrian Wells

164

8. Neurobiology Smit S. Sinha, Jan Mohlman, and Jack M. Gorman

187

xv

xvi

Contents

PART III. Assessment and Treatment 9. Assessment Cynthia L. Turk, Richard G. Heimberg, and Douglas S. Mennin

219

10. A Meta-Analytic Review of Cognitive-Behavioral Treatments Robert A. Gould, Steven A. Safren, David O’Neill Washington, and Michael W. Otto

248

11. Cognitive-Behavioral Therapy Robert L. Leahy

265

12. Supportive–Expressive Psychodynamic Therapy Paul Crits-Christoph, Mary Beth Connolly Gibbons, and Katherine Crits-Christoph

293

13. Integrative Psychotherapy Michelle G. Newman, Louis G. Castonguay, Thomas D. Borkovec, and Christine Molnar

320

14. Pharmacological Treatment R. Bruce Lydiard and Jeannine Monnier

351

PART IV. Special Populations 15. Children and Adolescents Anne Marie Albano and Sabine Hack

383

16. Older Adults J. Gayle Beck and Patricia M. Averill

409

Index

435

GENERALIZED ANXIETY DISORDER

PART I

Generalized Anxiety Disorder in the Clinic and in the Community

GAD IN Clinica l Presentation THE CLINICand AND Diag INnostic THE Features COMMUNITY

CHAPTER 1

Clinical Presentation and Diagnostic Features DOUGLAS S. MENNIN RICHARD G. HEIMBERG CYNTHIA L. TURK

Anxiety disorders have received increasing attention in the clinical psychological and psychiatric literature (Cox, Wessel, Norton, Swinson, & Direnfeld, 1995; Norton, Cox, Asmundson, & Maser, 1995). However, as Dugas (2000) notes, relatively few of these investigations have focused on generalized anxiety disorder (GAD). One explanation for this paucity of research is that it has taken a great deal of time for the diagnosis of GAD to establish validity, given major shifts in how it has been defined. This has presented a challenge to our understanding and treatment of the condition. However, this situation is now changing in a positive direction. This chapter reviews the development of GAD from a misunderstood and often maligned residual category to its current, more valid form, which has led to sizable advances in our understanding of its basic mechanisms. Specifically, this chapter covers (1) the history of GAD from Freud’s early writings to the present diagnostic system; (2) the definition and distinctness of the construct of worry, especially as it relates to GAD; and (3) criticisms of the current diagnostic system, including suggestions for new directions.

THE NOSOLOGICAL HISTORY OF GAD From Freud to DSM-II Discussion of a generalized anxiety syndrome can be seen as early as Freud’s writings on “anxiety neurosis” (e.g., Freud, 1920/1966). Freud dis3

4

GAD IN THE CLINIC AND IN THE COMMUNITY

tinguished between those who suffered from anxiety neurosis (which would include both GAD and panic disorder as currently defined) from those with “phobic neurosis” (which would include the current categories of agoraphobia, social phobia, and specific phobia). Rickels and Rynn (2001) note that Freud was the first to describe characteristics of worry, which he termed “anxious expectation.” However, unlike the current conceptualization that emphasizes the cognitive load of worrying, Freud viewed the condition as typically presenting as diffuse and somatic, appearing as “free-f loating anxiety” that often could not be readily identified by the patient. The first edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-I; American Psychiatric Association [APA], 1952) was heavily inf luenced by Freudian conceptualizations. Therefore, a patient with present-day GAD would probably have received a diagnosis of “anxiety reaction,” and later, in DSM-II (APA, 1968), “anxiety neurosis” (Rickels & Rynn, 2001). Neither early edition of the DSM separated generalized, chronic anxiety from more acute panic attacks. However, during the 1960s and 1970s, a number of events paved the way for the current form of diagnostic classification and, more specifically, for the appearance of GAD as a diagnostic category. First, Klein (1964) discovered that imipramine, a tricyclic antidepressant, was effective in treating acute panic but not more generalized anxiety—a discovery suggesting that these two types of anxiety may be distinct. Second, there was a growing movement away from theoretically based classification to more atheoretical, scientifically based diagnostic categories. Spitzer, Endicott, and Robins (1978) argued that much of the description provided in DSM-II was based on etiological conjecture rather than observable phenomena, and that without empirical validation of etiological factors, classification should be atheoretical and symptombased (as in other areas of medicine). Spitzer and colleagues developed a classification system of mental disorders to be used in research with the aim of providing empirical support for these categories. This system, the Research Diagnostic Criteria (RDC), was the precursor to the diagnostic system in place today, which is largely symptom-based, atheoretical, and increasingly specific in its approach to classification.

DSM-III and DSM-III-R DSM-III (APA, 1980) adopted the less theoretically biased and more empirical approach of the RDC, providing a specific descriptive nosology for the classification of anxiety disorders according to known or suspected differences in symptomatology, epidemiology, etiology, phenomenology, and treatment response. As a result, GAD and panic disorder/agoraphobia were separated. DSM-III also included phobic disorders, obsessive– compulsive disorder, and posttraumatic stress disorder. To meet criteria

Clinical Presentation and Diagnostic Features

5

for a DSM-III diagnosis of GAD, one needed to qualify for symptoms in at least three of the following four categories: (1) symptoms ref lecting tension, startle, and restlessness; (2) symptoms ref lecting autonomic hyperactivity; (3) symptoms ref lecting apprehensive expectation, including worry, fear, and rumination; and (4) symptoms ref lecting hypervigilance. Furthermore, one needed to have met these criteria for at least 1 month, during which time no other disorder could be present. Although this was the first instance in which GAD was considered a disorder in its own right, it would still be a long time until it would be accepted as an independent entity. DSM-III GAD was fraught with difficulty and was viewed with widespread skepticism in the psychiatric community. Early investigations into the reliability of GAD provided little support. In particular, interrater reliability was poor (kappa = .47; DiNardo, O’Brien, Barlow, & Waddell, 1983). This led some investigators to argue that GAD should not be considered a disorder in its own right (Breier, Charney, & Heninger, 1985). Poor reliability of the diagnosis of GAD stemmed largely from the structure of the DSM-III. DSM-III provided a hierarchical system of diagnosis that minimized concurrent diagnoses. For example, anxiety disorders were not diagnosed in the presence of pathology presumed to be more pervasive, such as depressive disorders or schizophrenia. Furthermore, GAD was to be considered only when all other anxiety disorders had been ruled out (Spitzer & Williams, 1985). Thus GAD was a poorly defined residual category and tended to be assigned to those patients who experienced significant anxiety but did not evidence phobic avoidance or panic attacks. Given the nature of the DSM-III GAD criteria, it was unlikely to demonstrate specificity. As Barlow (1988, p. 567) noted, “the category of GAD in DSM-III produced so much confusion that few clinicians or investigators could agree on individuals who would meet this definition.” As a result of difficulties in differential diagnosis (including, but not limited to, GAD), the hierarchical rules for the assignment of concurrent diagnoses were modified in DSM-III-R (APA, 1987). Although a more pervasive disorder would take precedence if it shared symptoms with a less pervasive disorder, additional Axis I diagnoses could be assigned if the focus of the secondary concern was not related to the first disorder. One caveat was that the diagnosis of GAD could not be given if it occurred solely during the course of a mood disorder. Although this revision maintained the need for differential diagnosis, strict hierarchical rules had been lifted. The result was that GAD could be considered a disorder in its own right, even when other anxiety disorders were present. In addition to the removal of hierarchical rules, the DSM-III-R diagnosis was more focused than its DSM-III predecessor. Whereas DSM-III largely retained Freud’s notion of “free-f loating anxiety,” DSM-III-R made

6

GAD IN THE CLINIC AND IN THE COMMUNITY

worry the central feature of GAD. In particular, to meet criteria for GAD, one had to demonstrate unrealistic and excessive worry or anxiety about two or more life circumstances (APA, 1987). In addition, anxiety and worry had to be present for at least 6 months (an increase from the less stringent criterion of 1 month in DSM-III). Data supported making worry the central feature of GAD. Sanderson and Barlow (1990) found that the presence of worry could be reliably diagnosed, with a kappa coefficient of .90. However, despite these changes, reliability for the DSM-III-R diagnosis of GAD remained moderate (kappa = .54; DiNardo, Moras, Barlow, Rapee, & Brown, 1993).

DSM-IV/DSM-IV-TR and ICD-10 One reason for the low reliability of DSM-III-R GAD may have been the fact that the associated symptom criteria remained diffuse. Although this list of symptoms was more detailed in DSM-III-R than in DSM-III (one needed to have 6 of 18 listed symptoms), the categories were still broad, including 4 symptoms of motor tension, 9 symptoms of autonomic hyperactivity, and 5 symptoms of vigilance and scanning. Most problematic was the inclusion of autonomic symptoms, which were found to be the least reliable aspect of the DSM-III-R diagnosis (Marten et al., 1993; Starcevic, Fallon, Uhlenhuth, & Pathak, 1994b). Marten and colleagues (1993) found that a reduced set of associated symptoms was more able to differentiate individuals with GAD from controls. As a result, DSM-IV adopted this reduced set of associated symptoms, which excluded autonomic symptoms but retained symptoms of tension and vigilance. In addition, the worry criteria were modified to ref lect a number of findings concerning components of the DSM-III-R diagnosis. First, investigations found that individuals with GAD were more likely than those without GAD to have difficulty controlling their worries (Craske, Rapee, Jackel, & Barlow, 1989), and difficulty controlling worry became one of the diagnostic criteria in DSM-IV. In contrast, content of worries appeared to have little ability to discriminate between individuals with GAD and controls (Abel & Borkovec, 1995). Specifically, whether worries were considered to be unrealistic or not had little discriminatory ability, and this was subsequently dropped from the criteria set. Furthermore, it was recognized that individuals with GAD could have pervasive worry regardless of how many different areas of their life were affected, and the requirement for two or more domains of worry was removed. Finally, overanxious disorder of childhood was subsumed under GAD (for a discussion of GAD in children, see Albano & Hack, Chapter 15, this volume). These changes resulted in a diagnosis that has continued to gain in reliability with current assessments (see Turk, Heimberg, & Mennin,

Clinical Presentation and Diagnostic Features

7

Chapter 9, this volume, for a discussion of the reliability of DSM-IV diagnostic assessments). In sum, the essential feature of DSM-IV GAD is excessive worry occurring more days than not for at least 6 months. The worry may concern a number of different domains or activities (e.g., work, finances, family, and health) and must be difficult to control. In addition, the worry or anxiety must be associated with at least three (one for children) of the following six symptoms: (1) restlessness, (2) fatigue, (3) impaired concentration, (4) irritability, (5) muscle tension, and (6) sleep disturbance. As in all DSM-IV diagnoses, the worry must lead to significant distress or impairment and must not be due to another Axis I disorder, to a general medical condition, or to a substance use disorder (APA, 1994). Furthermore, DSM-IV maintained the requirement that GAD must not be present solely during the course of a mood disorder. The latest version of the DSM is a text revision of the DSM-IV (DSM-IV-TR; APA, 2000). DSMIV-TR retained the DSM-IV diagnostic criteria set, but revised the text to ref lect more recent investigations into the nature of GAD. In particular, prevalence rates and genetic studies ref lecting the DSM-IV criteria were added. DSM-IV diagnostic criteria were also modified to be more compatible with the 10th revision of the International Classification of Diseases and Related Health Problems (ICD-10; World Health Organization, 1992). One study reported a kappa of .86 for agreement between interviewers based on DSM-IV and ICD-10 (Starcevic & Bogojevic, 1999) criteria for comorbid GAD in a sample of patients with primary panic disorder and agoraphobia. However, substantial differences remain. ICD-10 does not require worry to be experienced as excessive or difficult to control. Furthermore, ICD-10 has retained the requirement of symptoms of autonomic hyperactivity and continues to disallow the diagnosis of GAD if major depression, a phobic disorder, panic disorder, or obsessive–compulsive disorder is present (similar to DSM-III). Finally, there is no requirement that the worry or anxiety be clinically impairing, as in DSM-IV. Indeed, Slade and Andrews (2001) found that approximately 50% of patients who met ICD-10 criteria for GAD did not also meet DSM-IV criteria, demonstrating that these two classification systems may detect different patients.

COMORBIDITY WITH OTHER DISORDERS Although the DSM-IV/DSM-IV-TR diagnostic system allows more heterogeneous anxiety phenomenology within a diagnostic category, it also increases the likelihood that one will receive multiple diagnoses. “Comorbidity” refers to the occurrence of at least two different disorders in the same individual (Brown & Barlow, 1992). GAD has often been

8

GAD IN THE CLINIC AND IN THE COMMUNITY

found to be comorbid, with approximately 60–90% of patients meeting criteria for another disorder (e.g., Brawman-Mintzer et al., 1993; Brown, Barlow, & Liebowitz, 1994; Sanderson & Barlow, 1990). In fact, “pure” presentations of GAD are rare, especially when diagnoses are viewed longitudinally (e.g., Bruce, Machan, Dyck, & Keller, 2001). Some have argued that, given the high rates of comorbidity in the anxiety disorders, these categories should not be viewed as distinct entities, but rather as different expressions of the same underlying diathesis (Andrews, 1996; Clark, Watson, & Mineka, 1994). Achenbach (1995) explains that other explanations for comorbidity are that the disorders ref lect a higher-order taxon that includes both disorders; that the taxonomic borders between the disorders are not clearly specified; that the assessment materials do not clearly differentiate the disorders; or that a referral bias exists, in which cases with two disorders are more likely to be referred than individuals with only one disorder (commonly known as “Berkson’s bias”). Kessler, Walters, and Wittchen (Chapter 2, this volume) argue that Berkson’s bias may account for the high rates of comorbidity found in clinical samples. Nonetheless, it is clinically uncommon to see noncomorbid GAD; as such, the disorder often becomes difficult to detect when other disorders are present. In the remainder of this section, prevalence rates in clinical samples are reviewed (for rates in community samples, see Kessler et al., Chapter 2). GAD appears to be the most common additional diagnosis among patients with social phobia (Barlow, Blanchard, Vermilyea, Vermilyea, & DiNardo, 1986; deRuiter, Rijken, Garssen, van Schaik, & Kraaimaat, 1989; Sanderson, DiNardo, Rapee, & Barlow, 1990; Turner, Beidel, Borden, Stanley, & Jacob, 1991), with rates as high as 33% (Turner et al., 1991) and as low as 17% (Moras, unpublished manuscript, cited in Brown & Barlow, 1992) reported in the literature. Social phobia is also diagnosed frequently in patients with GAD, with rates between 23% (BrawmanMintzer et al., 1993) and 59% (Sanderson et al., 1990) reported. However, few studies have examined the characteristics of these disorders when they occur together. Mennin, Heimberg, and Jack (2000) examined GAD among patients with a principal diagnosis of social phobia. Patients with both social phobia and GAD demonstrated greater social anxiety and avoidance, general anxiety and negative mood, functional impairment, and overall psychopathology than patients with social phobia but without GAD. Early studies of patients with GAD showed that these patients often experience panic attacks (Barlow, 1988). Rates of panic disorder have been found to range from 11% to 27% in patients with GAD (BrawmanMintzer et al., 1993; Sanderson et al., 1990). Rates of GAD in patients with panic disorder have been found to range from 13% to 36% (Brown & Barlow, 1992; Sanderson et al., 1990). Obsessive–compulsive disorder can

Clinical Presentation and Diagnostic Features

9

also occur in patients with GAD, although reported rates have been comparatively low (1–11%; Brawman-Mintzer et al., 1993; Brown, Moras, Zinbarg, & Barlow, 1993; Sanderson et al., 1990). Rates of GAD in obsessive–compulsive disorder are slightly higher, ranging between 6.5% (Brown et al., 1993) and 20% (Abramowitz & Foa, 1998). Nevertheless, there has been considerable controversy surrounding the conceptual separation of these disorders and their cognitive components—namely, worries and obsessions (see further discussion of this issue below). GAD is often comorbid with mood disorders. Brawman-Mintzer and colleagues (1993) found that 42% of patients with GAD had a history of a past major depressive episode. GAD is also common in patients with major depressive disorder. Twenty percent of a sample of patients with major depressive disorder met criteria for GAD (Sanderson et al., 1990). DiNardo and Barlow (1990) reported unpublished findings showing a higher rate of GAD in major depression, with 45% meeting criteria. GAD has consistently been found to be the most common comorbid anxiety disorder among patients with dysthymic disorder. For example, Pini and colleagues (1997) found that 65% of patients with dysthymic disorder had GAD as a concurrent diagnosis, compared with 32% of patients with bipolar disorder and 37% of patients with major depressive disorder. Dysthymia is also a common secondary diagnosis in individuals with GAD, with rates as high as 50% reported (e.g., Garvey, Noyes, Anderson, & Cook, 1991). Studies have recently begun to examine the occurrence of Axis II disorders in patients with anxiety disorders. GAD has been found to have a stronger relationship with personality disorders than panic disorder or agoraphobia, and this relationship is strengthened when major depression is also present (Reich et al., 1994). Sanderson, Wetzler, Beck, and Betz (1994) found that 49% of patients with GAD met criteria for at least one comorbid personality disorder. Sanderson and Wetzler (1991) found avoidant personality disorder and dependent personality disorder to be the most common Axis II diagnoses in 32 patients with GAD.

DIFFERENTIAL DIAGNOSIS Do the high rates of comorbidity in clinical samples imply that GAD is not an entity in its own right? Should this disorder be subsumed under other diagnostic categories? Given the comorbidity rates presented above, one might expect GAD to have low levels of “specificity” (i.e., the extent to which its characteristics can be differentiated from those of other disorders). However, a number of studies have demonstrated the validity of the diagnosis by comparing symptom patterns with other diagnoses. One examination found evidence for the specificity of GAD by dem-

10

GAD IN THE CLINIC AND IN THE COMMUNITY

onstrating that the physical symptom criteria were endorsed at higher levels by these patients than by patients with other anxiety disorders (Ladouceur et al., 1999). Patients with GAD demonstrated higher levels of restlessness, insomnia, irritability, nausea, and headaches than patients with panic disorder (Nisita et al., 1990; Noyes et al., 1992). Conversely, patients with panic disorder had higher levels of trembling, rapid heartbeat, dizziness, faintness, dyspnea, chest pain, numbness, choking sensations, and derealization/depersonalization than patients with GAD (Nisita et al., 1990; Noyes et al., 1992). Noyes and colleagues (1992) concluded that GAD is characterized by central nervous system hyperarousal (e.g., muscle tension), whereas panic disorder is best characterized by autonomic arousal. (For a more detailed discussion of the distinctions between GAD and panic disorder, see Maser, 1998.) A number of studies also suggest that social phobia and GAD can be reliably distinguished. For example, patients with social phobia reported higher levels of sweating, faintness, and palpitations than patients with GAD, and patients with GAD reported greater fear of dying and dizziness than those with social phobia (Reich, Noyes, & Yates, 1988). In addition, those with social phobia reported less frequent headaches and a higher frequency of dyspnea than patients with GAD (Reich et al., 1988). Insomnia has also been found to be more common in patients with GAD than in those with social phobia (Versiani, Mundim, Nardi, & Liebowitz, 1988). Gross, Oei, and Evans (1989) found that, similar to patients with panic disorder, patients with social phobia are more characterized by autonomic arousal than patients with GAD. The research differentiating GAD from other anxiety disorders, such as panic disorder and social phobia, suggests that the removal of the autonomic hyperactivity symptom criteria in DSM-IV was warranted. A number of studies have found that, beyond worry, muscle tension and vigilance/scanning are the most commonly endorsed symptoms among patients with GAD (Abel & Borkovec, 1995; Brown, Marten, & Barlow, 1995; Marten et al., 1993; Starcevic, Fallon, Uhlenhuth, & Pathak, 1994a). In addition, these studies found that patients with GAD rarely endorsed autonomic symptoms. Laboratory studies using psychophysiological measurement have consistently found increases in electromyographic activity, suggesting elevated muscle tension (Hoehn-Saric, Hazlett, Pourmotabbed, & McLeod, 1997; Hoehn-Saric, McLeod, & Zimmerli, 1989). In particular, Hoehn-Saric and colleagues (1989) found that patients with GAD differed from controls on this measure of muscle tension, but not on physiological measures of autonomic activity. Finally, a number of studies have demonstrated that GAD could be distinguished from the other anxiety disorders based on presence of muscle tension (Brown, Antony, & Barlow, 1992; Joormann & Stöber, 1999). In fact, in a college sample, only muscle tension was uniquely associated with GAD (when the inf luence of depressive

Clinical Presentation and Diagnostic Features

11

symptoms was controlled for; Joormann & Stöber, 1999). In contrast, the relationship between difficulty concentrating and GAD was reduced to a nonsignificant level when depressive symptoms were controlled for, suggesting that this symptom has comparably low levels of specificity in contributing to the detection of GAD versus major depression. The results concerning muscle tension notwithstanding, the decision to make worry the central feature of GAD may be the strongest factor in the increase of reliability and validity of the DSM-IV diagnosis of GAD. In fact, Sanderson and Barlow (1990) found that excessive worry could be reliably diagnosed (kappa = .90). Although worry is common to all the anxiety disorders, it is present in greater amounts in GAD (Abel & Borkovec, 1995; Brown et al., 1993; Gross et al., 1989; Ladouceur et al., 1999). Borkovec (as reported in Brown et al., 1994) found that patients with GAD experienced a significantly greater degree of life interference due to their worry than patients with other anxiety disorders. Furthermore, there are substantial differences in the quality of this worry between patients with GAD and either patients with other disorders or nondisordered controls (Borkovec, Shadick, & Hopkins, 1991; Craske et al., 1989; Dugas et al., 1998; Starcevic, 1995). Barlow and colleagues (1986) first suggested that the content of the cognitions in patients with GAD was different from the focused thoughts of physical or mental catastrophe found in patients with panic disorder. They found that persons who reported worry about multiple life circumstances were more often diagnosed as having GAD. Persons who had focused anxiety about having a panic attack were more often diagnosed as having panic disorder. A more recent study by Breitholtz, Johansson, and Öst (1999) found that patients with panic disorder had significantly more cognitions concerning physical danger, whereas patients with GAD were characterized by cognitions focusing on interpersonal conf licts and worry about significant others. Sanderson and Barlow (1990) found that 91% of patients with GAD reported chronically worrying about minor events. This was significantly different from the 41% of patients with panic disorder who reported these same worries. This finding has been replicated in another study of GAD and panic disorder (Breitholtz et al., 1999), in a study comparing GAD and obsessive–compulsive disorder (Brown et al., 1993), and in a study comparing GAD and social phobia (Hoyer, Becker, & Roth, 2001). Dugas and colleagues (1998) found that patients with GAD worried more about the future than patients with other anxiety disorders. Ladouceur and colleagues (1999) found that patients with GAD scored higher than patients with other anxiety disorders on measures of problem-solving deficits and intolerance of uncertainty—two measures that have been shown to be related to worry. Also, Hoyer and colleagues (2001) found that higher frequency of worry, higher number of different worries, lower sub-

12

GAD IN THE CLINIC AND IN THE COMMUNITY

jective controllability, and distress during worry differentiated patients with GAD from both controls and patients with social phobia. Furthermore, the number of excessive worries was found to differentiate GAD from anxiety disorder not otherwise specified (the actual residual category in DSM-IV) in a Hispanic sample (Street et al., 1997). In contrast, one study found that levels of worry could not be differentiated between patients with GAD and major depression (Starcevic, 1995). However, Diefenbach and colleagues (2001) found that patients with GAD did differ from patients with a mood disorder in the content of their reported worries. Whereas the patients with GAD were more likely to focus their worries on loss of control, the depressed patients’ worry focused on an aimless future, lack of confidence, relationships, and financial concerns. This finding must be interpreted with caution, given that differences in worry content between disorders have often been found to be unreliable (see below; for a more in-depth discussion of the overlap of GAD and major depression, see Kessler et al., Chapter 2). Altogether, these findings demonstrate the importance of worry to GAD.

WORRY Definition and History Worry may characterize GAD, but how should worry be characterized? Worry is a universal experience, but a clear understanding of this phenomenon remains elusive. The word “worry” was first documented prior to the 12th century as a verb deriving from the Old English hunting term wyrgan, which meant “to kill by strangulation.” Wyrgan is etymologically related to the Lithuanian word vertzi, which means “to constrict.” These descriptions seem foreign to a modern understanding of the word “worry.” However, the first definitions of the verb form of “worry,” in the Merriam-Webster WWWebster Dictionary, are “to harass by tearing, biting, or snapping especially at the throat” and “to touch or disturb something repeatedly” (Merriam-Webster, 1999). Hallowell (1997) explains that just as a dog worries a bone by gnawing continually on it, people who worry chew and gnaw on a problem, devoting full attention to it. Adam Phillips (1993) discusses these early definitions of worry in his book On Kissing, Tickling, and Being Bored: Psychoanalytic Essays on the Unexamined Life. He notes that, in addition to their obvious violent tone, these definitions illustrate that worry was something done unto another rather than unto oneself. The psychological sense of “worry” does not appear in the vernacular until the early 19th century when the noun form of the word arose. The WWWebster Dictionary defines the noun form of “worry” as “a mental distress or agitation resulting from concern usually for something impending or anticipated” (Merriam-Webster, 1999). Phillips points

Clinical Presentation and Diagnostic Features

13

out that at a certain point in history, worrying became an act that people inf lict on themselves rather than others. He explains, “What was once done by the mouths of the rapacious, the desirous, is now done, often with a relentless weariness, by the minds of the troubled” (p. 51). Worry has only recently become a focus of psychological inquiry. In an attempt to define worry operationally, Challman (1974) echoed the focus on internal distress in the definitions given above when he described worry as “an unconsciously instigated mental activity that seeks to control fate through suffering” (p. 1141). As noted above, Sigmund Freud never wrote about worry per se, but he did write about the role of “anxious expectation” in the maintenance of anxiety. Davey and Tallis (1994) note that his account bears considerable resemblance to contemporary accounts of “anxious apprehension.” Barlow (1988) notes that a cardinal feature of anxious apprehension is “intense worry” (p. 250). Throughout the first half of the 20th century, few psychological investigators wrote about the nature of worry. Worry first received empirical attention in the test anxiety literature. Liebert and Morris (1967; Morris & Liebert, 1970) discovered that responses on a test anxiety questionnaire were composed of two distinct factors, which they labeled Worry and Emotionality. The Worry factor represented self-evaluative negative cognition about test performance. The Emotionality factor appeared to focus on awareness of feeling states and physiological activity. Worry has been found to have a stronger relationship than Emotionality to test performance (Deffenbacher & Deitz, 1978) and task-generated attentional interference (Deffenbacher, 1978). Beginning in the early 1980s, research devoted to worry in its own right emerged. In examining the psychological aspects of insomnia, Borkovec (1979) noted that many individuals who had difficulty sleeping were engaged in excessive, negatively laden cognitive activity that could best be described as worrying. Since that time, Borkovec and colleagues (for a review, see Borkovec, Alcaine, & Behar, Chapter 4, this volume) have extensively examined the nature of the worry process. Borkovec, Robinson, Pruzinsky, and DePree (1983) developed a tentative definition of worry that closely resembles the current definition of the noun form of “worry” in the WWWebster Dictionary: Worry is a chain of thoughts and images, negatively affect-laden and relatively uncontrollable; it represents an attempt to engage in mental problem-solving on an issue whose outcome is uncertain but contains the possibility of one or more negative outcomes; consequently, worry relates closely to the fear process. (p. 10)

Borkovec (1994) has extended this definition by stating that worry involves predominantly thought activity rather than imagery. He notes,

14

GAD IN THE CLINIC AND IN THE COMMUNITY

“Worry is a predominantly verbal–linguistic attempt to avoid future aversive events” (p. 7). Borkovec, Ray, and Stöber (1998) state that worry involves “talking to ourselves a lot about negative things, most often about negative events that we are afraid might happen in the future” (p. 562).

Differentiating Worry from Related Constructs Similar to the suggestions of the early test anxiety literature, the definitions provided by Borkovec and colleagues describe worry as a cognitive phenomenon distinct from the physiological components of anxiety. However, they do not clarify how worry is to be differentiated from cognitive components of anxiety or from other conceptually similar cognitive processes, such as obsessions. As a result, a number of investigations have sought to explicate the distinctness of worry from these related constructs.

Worry and Anxiety Prior to the study of worry in test anxiety, few investigators distinguished between the constructs of anxiety and worry (Breznitz, 1971). Clearly, a strong relationship between anxiety and worry exists. Had this not been the case, worry would not be considered central to the diagnosis of GAD. However, it remains unclear to what extent these constructs overlap, subsume each other, or are actually manifestations of the same underlying entity. Pruzinsky and Borkovec (1990) argue that distinguishing worry from anxiety is difficult because (1) worry may not in principle be distinguishable from anxiety except in artificial experimental situations; and (2) the elicitation of either cognitive or somatic components of an anxiety response in humans is likely to lead rather quickly to the other component. O’Neill (1985) has criticized the independent use of the construct of worry. He argues, rather, that worry should be viewed solely as the cognitive component of anxiety. Using data from worry research, O’Neill builds a case that worry and anxiety share the same behavioral concomitants, and that changes in associated phenomena are ref lected in similar changes in both worry and anxiety. Hence, in his view, the study of worry adds nothing substantial to the study of anxiety. Borkovec (1985) responds to this criticism by stating that, regardless of whether worry is or is not the cognitive component of anxiety, “we must next define the characteristics of that component and elucidate its functional relationships with other components and variables” (p. 481). Tallis, Eysenck, and Mathews (1991) argue that O’Neill presumes that the meaning of the term “anxiety” is widely agreed upon. In arguing to abandon the study of worry, O’Neill is suggesting that what is currently known about anxiety is satisfactory, is sufficient, and does not warrant further investigation into related

Clinical Presentation and Diagnostic Features

15

cognitive elements. Tallis and colleagues assert that the investigation of cognitive processes such as worry furthers our understanding of anxiety. Despite O’Neill’s arguments, a number of examinations have demonstrated that worry and anxiety can be reliably differentiated. An early examination of worry and fear found that in response to a stressor, fear was a sufficient but not a necessary condition for worry (Levy & Guttman, 1976). In other words, worry can occur in the absence of fear. Correlations between worry and anxiety have been found to be somewhat high, with values approximating .70 (Davey, 1993; Davey, Hampton, Farrell, & Davidson, 1992; Russell & Davey, 1993). However, worry and anxiety demonstrated a significantly lower correlation in patients with GAD (r = .18; Meyer, Miller, Metzger, & Borkovec, 1990). Worry and trait anxiety accounted for unique variance in problem-solving skills and coping styles in a sample of college students (Davey et al., 1992). In addition, whereas worry was more associated with problem-focused coping styles, trait anxiety was often associated with an avoidant style of coping and lack of problem-solving confidence. Induction of worry was found to produce a greater increase in negative cognitive intrusions than a neutral induction, whereas an anxiety induction did not (York, Borkovec, Vasey, & Stern, 1987). Stöber (1998) compared worried and anxious appraisals and found that, compared to worry, anxiety was more frequently focused on oneself than on others. Worry was more frequently associated with appraisals of others. Zebb and Beck (1998) found that worry was more highly associated with depression, confusion, lack of emotional control, and lack of control over problem solving than anxiety was. As described above, levels of worry cannot be differentiated between patients with GAD and patients with major depression (Starcevic, 1995). Furthermore, Andrews and Borkovec (1988) found that a worry induction produced both anxious and depressive affect. This effect was stronger than the effect of an anxiety induction on depressive affect or the effect of a depressive mood induction on anxious affect. These results suggest that worry may have a unique relationship to depressive affect beyond its common bond with anxiety.

Worry and Obsessions The delineation of worry as a cognitive process increases its distinction from anxiety, but blurs its distinction from other cognitive processes. Colloquially, “worrying” and “obsessing” are often considered synonymous. In fact, many clients seen in clinical practice describe their worrying as obsessing. However, obsessions are operationally defined quite differently in clinical psychology and psychiatry than in common usage. DSM-IV (APA, 1994) provides a detailed definition of obsessions as “recurrent and persistent thoughts, impulses, or images that are experienced, at some

16

GAD IN THE CLINIC AND IN THE COMMUNITY

time during the disturbance, as intrusive and inappropriate and that cause marked anxiety or distress” (p. 422). Investigations have shown that obsessions differ from worry in their antecedents (Craske et al., 1989), content (Borkovec et al., 1991), form (Borkovec & Inz, 1990), and acceptability (Roemer & Borkovec, 1993). Brown and colleagues (1993) found that patients with obsessive– compulsive disorder and patients with GAD were distinguishable on the basis of their self-reported obsessions and worries, respectively. Also, Abramowitz and Foa (1998) found that indecisiveness, pathological responsibility, and worry about everyday topics distinguished patients with obsessive–compulsive disorder who had comorbid GAD from those who did not. Examinations of worry and obsessions in nonclinical samples also lend support to their distinction. In a study by Coles, Mennin, and Heimberg (2001), items from measures of worry and obsession loaded on separate factors in an exploratory factor analysis. Wells and Morrison (1994), utilizing a diary methodology over a 2-week period, found worry to be reported as having more verbal content than obsessions, and obsessions to be more imagery-based. Worry was also rated as more realistic, more voluntary, harder to dismiss, more distracting, and of longer duration than obsessions. A recent set of studies (Langlois, Freeston, & Ladouceur, 2000a, 2000b) provided further evidence for the distinction between worries and obsessions. Results from these studies demonstrated that individuals differentiated these constructs on multiple dimensions, and that categorization errors (i.e., worry vs. obsession) were rare. A number of associated cognitive process variables have also been shown to have stronger relationships with either obsessions or worry. Dugas, Gosselin, and Ladouceur (2001) found that intolerance of uncertainty (see Dugas, Buhr, & Ladouceur, Chapter 6, this volume, for a discussion of this construct) was more associated with worry than with obsessions or compulsions. Furthermore, Coles and colleagues (2001) found that “thought–action fusion” (Rassin, Merckelbach, Muris, & Spaan, 1999), a cognitive process in which thought and deed are confused, could differentiate obsessions and worries. Specifically, thought–action fusion was strongly related to obsessions after controlling for the effect of worry. Hazlett-Stevens, Zucker, and Craske (2002) found that thought–action fusion was weakly correlated with worry but unrelated to a diagnosis of GAD. Overall, these findings demonstrate that worry and obsessions, although related processes, are conceptually distinct.

Worry and Rumination As described above, worry has been found to have a relationship to depression as well as to GAD (Diefenbach et al., 2001; Starcevic, 1995). Nolen-Hoeksema (1991) has found that rumination is related to the onset,

Clinical Presentation and Diagnostic Features

17

maintenance, and severity of depression. Nolen-Hoeksema (1998, p. 216) defines “depressive rumination,” a particular variety of rumination associated with depressed mood, as “focusing passively and repetitively on one’s symptoms of distress and the meaning of those symptoms without taking action to correct the problems one identifies.” The WWWebster Dictionary describes the etymology of “rumination” as deriving from the Latin ruminatus, which means “to chew repeatedly for an extended period of time” (Merriam-Webster, 1999). Similar to the derivation of worry, rumination refers to repetitive animal activity. In this case, cows are considered to ruminate food regurgitated from their multiple stomachs by chewing repeatedly. Research has linked depressive rumination to the onset of depression in nondepressed individuals (Just & Alloy, 1997), the exacerbation of depressed symptoms (Kuehner & Weber, 1999), and the increased chronicity of depressed symptoms (Nolen-Hoeksema, 2000). Stöber (2000) has argued that worry and rumination represent different aspects of the same underlying process. However, despite descriptive overlap between these constructs, few studies have directly compared them. A recent study (Fresco, Frankel, Mennin, Turk, & Heimberg, 2002) found that worry and rumination were correlated (r = .40) with each other, but were demonstrated to be distinctive via factor-analytic methods. However, these processes were not at all distinctive in their relationships to anxiety and depression. Segerstrom, Tsao, Alden, and Craske (2000) found that worry and rumination loaded together on a higherorder factor, which they characterized as a maladaptive response set. However, a recent intriguing investigation into numerous forms of repetitive thought replicated the aggregation of worry and rumination (in a multidimensional scaling procedure) in a “negative valence repetitive thought” dimension, but also found that they did not aggregate on a dimension of “purpose of repetitive thought” (Segerstrom, Stanton, Alden, & Shortridge, in press). A general characterization of these constructs may be too broad in scope, and thus may obscure important differences between worry and rumination. Further examination into the relationship between these processes is necessary before any definitive conclusions can be reached.

Characteristics of Pathological Worry When do worries become dysfunctional? Is worry always pathological? Bruhn (1990) offers that it is pathological when worry results in prolonged periods of introspection and social withdrawal; when the process of worry does not stimulate learning and extrapolation to other life situations; and when worry becomes so selfsatisfying that it promotes a pattern of helplessness. (p. 561)

18

GAD IN THE CLINIC AND IN THE COMMUNITY

Anxiety is often considered to be an evolutionarily adaptive phenomenon. The Yerkes–Dodson law illustrates that at moderate levels, anxiety can improve performance. Indeed, it is difficult to imagine life without anxiety. Can the same be said of worry? Is it normal to worry? Worry may exist in nonpathological forms. Davey (1994) has criticized the definitions of worry reviewed above, stating that they focus on pathological forms of worry without addressing worry as a normal process. Hallowell (1997) suggests that worry can be helpful in motivating one to seek important goals. This view is ref lected in an early account of worry by Janis (1958), who argued that when a situation contains a realistic difficulty, worry may help one cope in the long term. But worry is not always a response to realistic threat. In addition, worry can often become aversive and uncontrollable in itself. How should this pathological type of worry be differentiated from a normal adaptive response? Some investigators have sought to differentiate the content of pathological worry from that of worry in nondisordered populations. Borkovec and colleagues (1983) examined domains of worry among college students. Whereas academic and interpersonal worries were rated as most frequent, concerns about physical harm were rated as least frequent. Wisocki (1988) found that among elderly individuals, health concerns were rated most highly, followed by interpersonal relations and finances. Craske and colleagues (1989) compared the content of worries in patients with GAD and nondisordered controls. They found that illness/health (31%) and family-related (28%) worries were most frequent, and worries about finances (3%) least frequent, in the group with GAD. The control participants in this study also reported worries about family-related issues (26%) as important, but rated work/school domains (30%) most frequently the objects of worry. For this group, illness/health worries (2%) were infrequent. Sanderson and Barlow (1990) also found that familyrelated concerns (79%) were frequent in patients with GAD. However, whereas Craske and colleagues found financial worries to be relatively infrequent in patients with GAD, Sanderson and Barlow (1990) found financial worries to be quite common (50%). Borkovec and colleagues (1991) have also reported family and interpersonal concerns (31%) to be frequent in patients with GAD. However, control participants also had high levels of family and interpersonal worries (44%) and had higher levels of illness/health worries (25%) than the group with GAD (3%). Finally, Roemer, Molina, and Borkovec (1997) found that patients with GAD could only be differentiated from control participants by the presence of worry over minor matters. The results from these studies suggest that examinations of the content of worries provide little discrimination between individuals with pathological and nonpathological worries. In fact, domains of worry were found to be inconsistent within groups with GAD across studies.

Clinical Presentation and Diagnostic Features

19

Patients with GAD and individuals with nonpathological worries do not seem to differ in the content of their worries. However, these groups may differ in the severity and pervasiveness of their worries. In fact, Mennin, Fresco, and Heimberg (1999) found that a dimensional measure of pathological worry had better sensitivity and specificity in identifying GAD than a measure of content domains of worry. This suggests that pathological worry may be a dimensionally more severe form of nonpathological worry. Indeed, one investigation utilizing taxometric analysis found that pathological worry and “normal” worry were actually dimensionally related rather than categorically distinct (Ruscio, Borkovec, & Ruscio, 2001). Furthermore, Craske and colleagues (1989) found that patients with GAD spent more time during the day worrying than did control participants. In addition, these patients were significantly less likely than control participants to recognize the precipitant of the worry and were significantly more likely to rate their worries as uncontrollable than control participants. Taken together, the research on worry demonstrates a specific relationship between pervasive levels of worry and GAD, and thus provides further validation of the current diagnostic classification.

CONCLUSIONS: LOOKING AHEAD TO DSM-V A number of investigators continue to examine the viability of the DSMIV-TR diagnosis and to determine what changes (if any) are necessary for the next edition of the diagnostic nomenclature, DSM-V. Some investigators have questioned the central focus on worry in the diagnosis of GAD (e.g., Rickels & Rynn, 2001). Bienvenu, Nestadt, and Eaton (1998) found that the associated six-symptom criteria set provided a better differentation between individuals with and without GAD than the worry criteria did. However, it is important to note that this study utilized DSM-III-R criteria rather than the more refined DSM-IV criteria. Nonetheless, Rickels and Rynn (2001) argue that there is too great a focus on the worry component of GAD. They argue that chronic worrying does not always lead to a severe anxiety syndrome, given that many individuals with chronic worries do not meet criteria for GAD (Ruscio, 2002). Other possible variables that have demonstrated a role in GAD beyond worry are beginning to receive attention. Intolerance of uncertainty has been shown to be a cognitive construct that is independent from worry and has discriminated GAD from other anxiety disorders (see Dugas et al., Chapter 6, for a discussion of this construct and related findings). Furthermore, Wells and Carter (2001) found that metacognitive beliefs about the detrimental nature of worry (i.e., worry about worrying) distinguished patients with GAD from those with other anxiety disorders when worry itself was controlled for. Finally, recent evidence suggests that

20

GAD IN THE CLINIC AND IN THE COMMUNITY

emotion regulation deficits predict a diagnosis of GAD beyond the effects of worry, anxiety, and associated depression (Mennin, Heimberg, Turk, & Fresco, 2003). In particular, persons with GAD may experience emotions with greater intensity than persons with social phobia do (Turk, Heimberg, Luterek, Mennin, & Fresco, in press). Similarly, Roemer, Salters, Raffa, and Orsillo (in press) found unique associations between emotional avoidance and the severity of GAD when worry was controlled for. Another criticism of the DSM-IV/DSM-IV-TR diagnosis of GAD is its placement on Axis I. A number of investigators have suggested that GAD may be better conceptualized as a personality or temperament disorder, given its chronicity, pervasiveness, and resistance to treatment (Akiskal, 1998; Rickels & Rynn, 2001; Sanderson & Wetzler, 1991). Another challenge to the current diagnostic criteria is ref lected in studies finding that a 6-month duration requirement may be overly restrictive and miss significant cases of GAD (Maier et al., 2000; see Kessler et al., Chapter 2, this volume). A recent paper also calls into question the mood disorder exclusionary criterion (in which one may not diagnose GAD if it fully overlaps with a mood disorder). Zimmerman and Chelminski (2003) compared patients with major depressive disorder and nonoverlapping GAD; patients with major depressive disorder and fully overlapping GAD (these patients would normally not receive a diagnosis of GAD, given the overlap); and patients without GAD or major depressive disorder. Both groups of depressed patients were found to have higher levels of suicidal ideation, social impairment, anxiety disorder comorbidity, eating disorders, somatoform disorders, and pathological worry, as well as a greater likelihood of a first-degree relative aff licted with GAD, than control patients. However, supporting the removal of this exclusionary criterion, the groups with GAD (i.e., overlapping vs. nonoverlapping) did not differ from each other. Finally, Barlow and Wincze (1998) have suggested that the somatic symptom criteria set may require modification. Citing Brown and colleagues (1995), who found that many patients without GAD endorsed three of the six associated symptoms, they suggest raising the required number of symptoms to four. A number of other associated somatic symptoms, which are typically found in patients with GAD in clinical settings, may also be found to have a more specific relationship to GAD (e.g., tension headaches). In summary, GAD has undergone major changes since its inception as a diagnosis over 20 years ago. Initially a “wastebasket” category that most mental health professionals viewed with a jaundiced eye, GAD is now widely considered a valid diagnostic category with significant associated disability. The reliability of the GAD diagnosis has improved substantially over this time, but continues to need improvement. More specific

Clinical Presentation and Diagnostic Features

21

delineation of the central features of the disorder and proposed criteria are necessary as we begin to conceptualize changes for DSM-V. Consensus among researchers in operational definitions of each GAD criterion would further increases in reliability (given GAD’s often variable and confusing diagnostic history). Nonetheless, in spite of high comorbidity rates, GAD can be reliably distinguished from other anxiety disorders. In particular, muscle tension and pervasive, excessive, and uncontrollable worry appear to be the most specific features of the condition. Finally, the study of worry has advanced our understanding of GAD. However, there are exciting new possible directions to expand our understanding of GAD, which may take us beyond the current focus on worry into the role of such factors as emotion dysregulation and interpersonal dysfunction. New investigations into these unexplored areas may help improve our understanding and treatment of GAD, thus providing further validation for this misunderstood diagnosis. REFERENCES Abel, J. L., & Borkovec, T. D. (1995). Generalizability of DSM-III-R generalized anxiety disorders to proposed DSM-IV criteria and cross-validation of proposed changes. Journal of Anxiety Disorders, 9, 303–315. Abramowitz, J. S., & Foa, E. B. (1998). Worries and obsessions in individuals with obsessive–compulsive disorder with and without comorbid generalized anxiety disorder. Behaviour Research and Therapy, 36, 695–700. Achenbach, T. (1995). Diagnosis, assessment, and comorbidity in psychosocial treatment research. Journal of Abnormal Child Psychology, 23, 45–65. Akiskal, H. S. (1998). Toward a definition of generalized anxiety disorder as an anxious temperament type. Acta Psychiatrica Scandinavica, 393(Suppl.), 66–73. American Psychiatric Association (APA). (1952). Diagnostic and statistical manual of mental disorders. Washington, DC: Author. American Psychiatric Association (APA). (1968). Diagnostic and statistical manual of mental disorders (2nd ed.). Washington, DC: Author. American Psychiatric Association (APA). (1980). Diagnostic and statistical manual of mental disorders (3rd ed.). Washington, DC: Author. American Psychiatric Association (APA). (1987). Diagnostic and statistical manual of mental disorders (3rd ed., rev.). Washington, DC: Author. American Psychiatric Association (APA). (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author. American Psychiatric Association (APA). (2000). Diagnostic and statistical manual of mental disorders (4th ed., text rev.). Washington, DC: Author. Andrews, G. (1996). Comorbidity in neurotic disorders: The similarities are more important than the differences. In R. M. Rapee (Ed.), Current controversies in the anxiety disorders (pp. 3–20). New York: Guilford Press. Andrews, V. H., & Borkovec, T. D. (1988). The differential effects of inductions of worry, somatic anxiety, and depression on emotional experience. Journal of Behavior Therapy and Experimental Psychiatry, 19, 21–26. Barlow, D. H. (1988). Anxiety and its disorders. New York: Guilford Press.

22

GAD IN THE CLINIC AND IN THE COMMUNITY

Barlow, D. H., Blanchard, E. B., Vermilyea, J. A., Vermilyea, B. B., & DiNardo, P. A. (1986). Generalized anxiety and generalized anxiety disorder: Description and reconceptualization. American Journal of Psychiatry, 143, 40–44. Barlow, D. H., & Wincze, J. (1998). DSM-IV and beyond: What is generalized anxiety disorder? Acta Psychiatrica Scandinavica, 393(Suppl.), 23–29. Bienvenu, O. J., Nestadt, G., & Eaton, W. W. (1998). Characterizing generalized anxiety: Temporal and symptomatic thresholds. Journal of Nervous and Mental Disease, 186, 51–56. Borkovec, T. D. (1979). Pseudo(experiential)-insomnia and idiopathic(objective) insomnia: Theoretical and therapeutic issues. Advances in Behaviour Research and Therapy, 2, 27–55. Borkovec, T. D. (1985). Worry: A potentially valuable concept. Behaviour Research and Therapy, 23, 481–482. Borkovec, T. D. (1994). The nature, functions, and origins of worry. In G. C. L. Davey & F. Tallis (Eds.), Worrying: Perspectives on theory, assessment, and treatment (pp. 5– 33). Chichester, UK: Wiley. Borkovec, T. D., & Inz, J. (1990). The nature of worry in generalized anxiety disorder: A predominance of thought activity. Behaviour Research and Therapy, 28, 153–158. Borkovec, T. D., Ray, W. J., & Stöber, J. (1998). Worry: A cognitive phenomenon intimately linked to affective, physiological, and interpersonal behavioral processes. Cognitive Therapy and Research, 22, 561–576. Borkovec, T. D., Robinson, E., Pruzinsky, T., & DePree, J. A. (1983). Preliminary exploration of worry: Some characteristics and processes. Behaviour Research and Therapy, 21, 9–16. Borkovec, T. D., Shadick, R. N., & Hopkins, M. (1991). The nature of normal and pathological worry. In R. M. Rapee & D. H. Barlow (Eds.), Chronic anxiety: Generalized anxiety disorder and mixed anxiety–depression (pp. 29–51). New York: Guilford Press. Brawman-Mintzer, O., Lydiard, R. B., Emmanuel, N., Payeur, R., Johnson, M., Roberts, J., Jarrell, M. P., & Ballenger, J. C. (1993). Psychiatric comorbidity in patients with generalized anxiety disorder. American Journal of Psychiatry, 150, 1216–1218. Breier, A., Charney, D. S., & Heninger, G. R. (1985). The diagnostic validity of anxiety disorders and their relationship to depressive illness. American Journal of Psychiatry, 142, 787–797. Breitholtz, E., Johansson, B., & Öst, L. G. (1999). Cognitions in generalized anxiety disorder and panic disorder patients: A prospective approach. Behaviour Research and Therapy, 37, 533–544. Breznitz, S. (1971). A study of worrying. British Journal of Social and Clinical Psychology, 10, 271–279. Brown, T. A., Antony, M. M., & Barlow, D. H. (1992). Psychometric properties of the Penn State Worry Questionnaire in a clinical anxiety disorders sample. Behaviour Research and Therapy, 30, 33–37. Brown, T. A., & Barlow, D. H. (1992). Comorbidity among anxiety disorders: Implications for treatment and DSM-IV. Journal of Consulting and Clinical Psychology, 60, 835–844. Brown, T. A., Barlow, D. H., & Liebowitz, M. R. (1994). The empirical basis of generalized anxiety disorder. American Journal of Psychiatry, 151, 1272–1280. Brown, T. A., Marten, P. A., & Barlow, D. H. (1995). Discriminant validity of the symptoms constituting the DSM-III-R and DSM-IV associated symptom criterion of generalized anxiety disorder. Journal of Anxiety Disorders, 9, 317–328.

Clinical Presentation and Diagnostic Features

23

Brown, T. A., Moras, K., Zinbarg, R. E., & Barlow, D. H. (1993). Diagnostic and symptom distinguishability of generalized anxiety disorder and obsessive–compulsive disorder. Behavior Therapy, 24, 227–240. Bruce, S. E., Machan, J. T., Dyck, I., & Keller, M. B. (2001). Infrequency of “pure” GAD: Impact of psychiatric comorbidity on clinical course. Depression and Anxiety, 14, 219–225. Bruhn, J. G. (1990). The two sides of worry. Southern Medical Journal, 83, 557–562. Challman, A. (1974). The empirical nature of worry. American Journal of Psychiatry, 131, 1140–1141. Clark, L. A., Watson, D., & Mineka, S. (1994). Temperament, personality, and the mood and anxiety disorders. Journal of Abnormal Psychology, 103, 103–116. Coles, M., Mennin, D. S., & Heimberg, R. G. (2001). Differentiating obsessions and worries: The role of thought–action fusion. Behaviour Research and Therapy, 39, 947–959. Cox, B. J., Wessel, I., Norton, G. R., Swinson, R. P., & Direnfeld, D. M. (1995). Publication trends in anxiety disorders research: 1990–1992. Journal of Anxiety Disorders, 9, 531–538. Craske, M. G., Rapee, R. M., Jackel, L., & Barlow, D. H. (1989). Qualitative dimensions of worry in DSM-III-R generalized anxiety disorder subjects and nonanxious controls. Behaviour Research and Therapy, 27, 397–402. Davey, G. C. L. (1993). A comparison of three worry questionnaires. Behaviour Research and Therapy, 31, 51–56. Davey, G. C. L. (1994). Pathological worrying as exacerbated problem-solving. In G. C. L. Davey & F. Tallis (Eds.), Worrying: Perspectives on theory, assessment, and treatment (pp. 35–59). Chichester, UK: Wiley. Davey, G. C. L., Hampton, J., Farrell, J., & Davidson, S. (1992). Some characteristics of worrying: Evidence for worrying and anxiety as separate constructs. Personality and Individual Differences, 13, 133–147. Davey, G. C. L., & Tallis, F. (Eds.). (1994). Worrying: Perspectives on theory, assessment, and treatment. Chichester, UK: Wiley. Deffenbacher, J. L. (1978). Worry, emotionality, and task-generated interference in test anxiety: An empirical test of attentional theory. Journal of Educational Psychology, 70, 248–254. Deffenbacher, J. L., & Deitz, S. R. (1978). Effects of test anxiety on performance, worry, and emotionality in naturally occurring exams. Psychology in the Schools, 15, 446– 450. deRuiter, C., Rijken, H., Garssen, B., van Schaik, A., & Kraaimaat, F. (1989). Comorbidity among the anxiety disorders. Journal of Anxiety Disorders, 3, 57–68. Diefenbach, G. J., McCarthy-Larzelere, M. E., Williamson, D. A., Mathews, A., Manguno-Mire, G. M., & Bentz, B. G. (2001). Anxiety, depression, and the content of worries. Depression and Anxiety, 14, 247–250. DiNardo, P. A., & Barlow, D. H. (1990). Syndrome and symptom comorbidity in the anxiety disorders. In J. D. Maser & C. R. Cloninger (Eds.), Comorbidity in anxiety and mood disorders (pp. 205–230). Washington, DC: American Psychiatric Press. DiNardo, P. A., Moras, K., Barlow, D. H., Rapee, R. M., & Brown, T. A. (1993). Reliability of DSM-III-R anxiety disorder categories: Using the Anxiety Disorders Interview Schedule—Revised (ADIS-R). Archives of General Psychiatry, 50, 251–256. DiNardo, P. A., O’Brien, G. T., Barlow, D. H., & Waddell, M. T. (1983). Reliability of DSM-III anxiety disorder categories using a new structured interview. Archives of General Psychiatry, 40, 1070–1074.

24

GAD IN THE CLINIC AND IN THE COMMUNITY

Dugas, M. J. (2000). Generalized anxiety disorder publications: So where do we stand? Journal of Anxiety Disorders, 14, 31–40. Dugas, M. J., Freeston, M. H., Ladouceur, R., Rheaume, J., Provencher, M., & Boisvert, J.-M. (1998). Worry themes in primary GAD, secondary GAD, and other anxiety disorders. Journal of Anxiety Disorders, 12, 253–261. Dugas, M. J., Gosselin, P., & Ladouceur, R. (2001). Intolerance of uncertainty and worry: Investigating specificity in a nonclinical sample. Cognitive Therapy and Research, 25, 551–558. Fresco, D. M., Frankel, A. N., Mennin, D. S., Turk, C. L., & Heimberg, R. G. (2002). Distinct and overlapping features of rumination and worry: The relationship of cognitive production to negative affective states. Cognitive Therapy and Research, 26, 179–188. Freud, S. (1966). Anxiety. In J. Strachey (Ed. & Trans.), Introductory lectures on psychoanalysis (pp. 487–511). New York: Norton. (Original work published 1920) Garvey, M. J., Noyes, R., Anderson, D., & Cook, B. (1991). Examination of comorbid anxiety in psychiatric inpatients. Comprehensive Psychiatry, 32, 277–282. Gross, P. R., Oei, T. P., & Evans, L. (1989). Generalized anxiety symptoms in phobic disorders and anxiety states: A test of the worry hypothesis. Journal of Anxiety Disorders, 3, 159–169. Hallowell, E. M. (1997). Worry: Hope and help for a common condition. New York: Ballantine Books. Hazlett-Stevens, H., Zucker, B. G., & Craske, M. G. (2002). The relationship of thoughtaction fusion to pathological worry and generalized anxiety disorder. Behaviour Research and Therapy, 40, 1199–1204. Hoehn-Saric, R., Hazlett, R. L., Pourmotabbed, T., & McLeod, D. R. (1997). Does muscle tension ref lect arousal?: Relationship between electromyographic and electroencephalographic recordings. Psychiatry Research, 71, 49–55. Hoehn-Saric, R., McLeod, D. R., & Zimmerli, W. D. (1989). Somatic manifestations in women with generalized anxiety disorder: Psychophysiological responses to psychological stress. Archives of General Psychiatry, 46, 1113–1119. Hoyer, J., Becker, E. S., & Roth, W. T. (2001). Characteristics of worry in GAD patients, social phobics, and controls. Depression and Anxiety, 13, 89–96. Janis, I. (1958). Psychological stress. New York: Wiley. Joormann, J., & Stöber, J. (1999). Somatic symptoms of generalized anxiety disorder for the DSM-IV: Associations with pathological worry and depression symptoms in a nonclinical sample. Journal of Anxiety Disorders, 13, 491–503. Just, N., & Alloy, L. B. (1997). The response styles theory of depression: Tests and an extension of the theory. Journal of Abnormal Psychology, 106, 221–229. Klein, D. (1964). Delineation of two drug-responsive anxiety syndromes. Psychopharmacologica, 5, 397–408. Kuehner, C., & Weber, I. (1999). Responses to depression in unipolar depressed patients: An investigation of Nolen-Hoeksema’s response styles theory. Psychological Medicine, 29, 1323–1333. Ladouceur, R., Dugas, M. J., Freeston, M. H., Rhéaume, J., Blais, F., Boisvert, J. M., & Thibodeau, N. (1999). Specificity of generalized anxiety disorder symptoms and processes. Behavior Therapy, 30, 191–208. Langlois, F., Freeston, M. H., & Ladouceur, R. (2000a). Differences and similarities between obsessive intrusive thoughts and worry in a non-clinical population: Study 1. Behaviour Research and Therapy, 38, 157–173.

Clinical Presentation and Diagnostic Features

25

Langlois, F., Freeston, M. H., & Ladouceur, R. (2000b). Differences and similarities between obsessive intrusive thoughts and worry in a non-clinical population: Study 2. Behaviour Research and Therapy, 38, 175–189. Levy, S., & Guttman, L. (1976). Worry, fear, and concern differentiated. Israel Annals of Psychiatry and Related Disciplines, 14, 211–228. Liebert, R. M., & Morris, L. W. (1967). Cognitive and emotional components of test anxiety: A distinction and some initial data. Psychological Reports, 20, 975–978. Maier, W., Gransicke, M., Freyberger, H. J., Linz, M., Heun, R., & Lecrubier, Y. (2000). Generalized anxiety disorder (ICD-10) in primary care from a cross-cultural perspective: A valid diagnostic entity? Acta Psychiatrica Scandinavica, 101, 29–36. Marten, P. A., Brown, T. A., Barlow, D. H., Borkovec, T. D., Shear, M. K., & Lydiard, R. B. (1993). Evaluation of the ratings comprising the associated symptom criterion of DSM-III-R generalized anxiety disorder. Journal of Nervous and Mental Disease, 181, 676–682. Maser, J. D. (1998). Generalized anxiety disorder and its comorbidities: Disputes at the boundaries. Acta Psychiatrica Scandinavica, 393(Suppl.), 12–22. Mennin, D. S., Fresco, D. M., & Heimberg, R. G. (1999, November). Determining when worry becomes clinically significant: A receiver operating characteristic (ROC) analysis approach. Paper presented at the annual meeting of the Association for the Advancement of Behavior Therapy, Toronto. Mennin, D. S., Heimberg, R. G., & Jack, M. S. (2000). Comorbid generalized anxiety disorder in primary social phobia: Symptom severity, functional impairment, and treatment response. Journal of Anxiety Disorders, 14, 325–343. Mennin, D. S., Heimberg, R. G., Turk, C. L., & Fresco, D. M. (2003). Preliminary evidence for an emotion regulation deficit model of generalized anxiety disorder: Testing a theoretical model. Manuscript submitted for publication. Merriam-Webster. (1999). WWWebster Dictionary [Online]. Available: http://www.mw.com Meyer, T. J., Miller, M. L., Metzger, R. L., & Borkovec, T. D. (1990). Development and validation of the Penn State Worry Questionnaire. Behaviour Research and Therapy, 28, 487–495. Morris, L. W., & Liebert, R. M. (1970). Relationship of cognitive and emotional components of test anxiety to physiological arousal and academic performance. Journal of Consulting and Clinical Psychology, 35, 332–337. Nisita, C., Petracca, A., Akiskal, H. S., Galli, L., Gepponi, I., & Cassano, G. B. (1990). Delimitation of generalized anxiety disorder: Clinical comparisons with panic and major depressive disorders. Comprehensive Psychiatry, 31, 409–415. Nolen-Hoeksema, S. (1991). Responses to depression and their effects on the duration of depressive episodes. Journal of Abnormal Psychology, 100, 569–582. Nolen-Hoeksema, S. (1998). The other end of the continuum: The costs of rumination. Psychological Inquiry, 9, 216–219. Nolen-Hoeksema, S. (2000). Further evidence for the role of psychosocial factors in depression chronicity. Clinical Psychology: Science and Practice, 7, 224–227. Norton, G. R., Cox, B. J., Asmundson, G. J. G., & Maser, J. D. (1995). The growth of research on anxiety disorders in the 1980s. Journal of Anxiety Disorders, 9, 75–85. Noyes, R., Jr., Woodman, C., Garvey, M. J., Cook, B. L., Suelzer, M., Clancy, J., & Anderson, D. J. (1992). Generalized anxiety disorder vs. panic disorder: Distinguishing characteristics and patterns of comorbidity. Journal of Nervous and Mental Disease, 180, 369–379.

26

GAD IN THE CLINIC AND IN THE COMMUNITY

O’Neill, G. W. (1985). Is worry a valuable concept? Behaviour Research and Therapy, 23, 479–480. Phillips, A. (1993). On kissing, tickling, and being bored: Psychoanalytic essays on the unexamined life. Cambridge, MA: Harvard University Press. Pini, S., Cassano, G. B., Simonini, E., Savino, M., Russo, A., & Montgomery, S. A. (1997). Prevalence of anxiety disorders comorbidity in bipolar depression, unipolar depression and dysthymia. Journal of Affective Disorders, 42, 145–153. Pruzinsky, T., & Borkovec, T. D. (1990). Cognitive and personality characteristics of worriers. Behaviour Research and Therapy, 28, 507–512. Rassin, E., Merckelbach, H., Muris, P., & Spaan, V. (1999). Thought—action fusion as a causal factor in the development of intrusions. Behaviour Research and Therapy, 37, 231–237. Reich, J. H., Noyes, R., & Yates, W. (1988). Anxiety symptoms distinguishing social phobia from panic and generalized anxiety disorders. Journal of Nervous and Mental Disease, 176, 510–513. Reich, J. H., Perry, J. C., Shera, D., Dyck, I., Vasile, R., Goisman, R. M., Rodriguez-Villa, F., Massion, A. O., & Keller, M. (1994). Comparison of personality disorders in different anxiety disorder diagnoses: Panic, agoraphobia, generalized anxiety, and social phobia. Annals of Clinical Psychiatry, 6, 125–134. Rickels, K., & Rynn, M. A. (2001). What is generalized anxiety disorder? Journal of Clinical Psychiatry, 62(Suppl. 11), 4–12. Roemer, L., & Borkovec, T. D. (1993). Worry: Unwanted cognitive activity that controls unwanted somatic experience. In D. M. Wegner & J. W. Pennebaker (Eds.), Handbook of mental control (pp. 220–238). Englewood Cliffs, NJ: Prentice-Hall. Roemer, L., Molina, S., & Borkovec, T. D. (1997). An investigation of worry content among generally anxious individuals. Journal of Nervous and Mental Disease, 185, 314–319. Roemer, L., Salters, K., Raffa, S., & Orsillo, S. M. (in press). Fear and avoidance of internal experiences in GAD: Preliminary tests of a conceptual model. Cognitive Therapy and Research. Ruscio, A. M. (2002). Delimiting the boundaries of generalized anxiety disorder: Differentiating high worriers with and without GAD. Journal of Anxiety Disorders, 16, 377–400. Ruscio, A. M., Borkovec, T. D., & Ruscio, J. (2001). A taxometric investigation of the latent structure of worry. Journal of Abnormal Psychology, 110, 413–422. Russell, M., & Davey, G. C. (1993). The relationship between life event measures and anxiety and its cognitive correlates. Personality and Individual Differences, 14, 317– 322. Sanderson, W. C., & Barlow, D. H. (1990). A description of patients diagnosed with DSM-III-R generalized anxiety disorder. Journal of Nervous and Mental Disease, 178, 588–591. Sanderson, W. C., DiNardo, P. A., Rapee, R. M., & Barlow, D. H. (1990). Syndrome comorbidity in patients diagnosed with a DSM-III-R anxiety disorder. Journal of Abnormal Psychology, 99, 308–312. Sanderson, W. C., & Wetzler, S. (1991). Chronic anxiety and generalized anxiety disorder: Issues in comorbidity. In R. M. Rapee & D. H. Barlow (Eds.), Chronic anxiety: Generalized anxiety disorder and mixed anxiety–depression (pp. 119–135). New York: Guilford Press.

Clinical Presentation and Diagnostic Features

27

Sanderson, W. C., Wetzler, S., Beck, A. T., & Betz, F. (1994). Prevalence of personality disorders among patients with anxiety disorders. Psychiatry Research, 51, 167–174. Segerstrom, S. C., Stanton, A. L., Alden, L. E., & Shortridge, B. E. (in press). A multidimensional structure for repetitive thought: What’s on your mind, and how, and how much? Journal of Personality and Social Psychology. Segerstrom, S. C., Tsao, J. C. I., Alden, L. E., & Craske, M. G. (2000). Worry and rumination: Repetitive thought as a concomitant and predictor of negative mood. Cognitive Therapy and Research, 24, 671–688. Slade, T., & Andrews, G. (2001). DSM-IV and ICD-10 generalized anxiety disorder: Discrepant diagnoses and associated disability. Social Psychiatry and Psychiatric Epidemiology, 36, 45–51. Spitzer, R. L., Endicott, J., & Robins, E. (1978). Research Diagnostic Criteria: Rationale and reliability. Archives of General Psychiatry, 35, 773–782. Spitzer, R. L., & Williams, J. B. W. (1985). Proposed revisions in the DSM-III classification of anxiety disorders based on research and clinical experience. In J. D. Maser & A. H. Tuma (Eds.), Anxiety and the anxiety disorders (pp. 759–773). Hillsdale, NJ: Erlbaum. Starcevic, V. (1995). Pathological worry in major depression: A preliminary report. Behaviour Research and Therapy, 33, 55–56. Starcevic, V., & Bogojevic, G. (1999). The concept of generalized anxiety disorder: Between the too narrow and too wide diagnostic criteria. Psychopathology, 32, 5– 11. Starcevic, V., Fallon, S., Uhlenhuth, E. H., & Pathak, D. (1994a). Comorbidity rates do not support distinction between panic disorder and generalized anxiety disorder. Psychopathology, 27, 269–272. Starcevic, V., Fallon, S., Uhlenhuth, E. H., & Pathak, D. (1994b). Generalized anxiety disorder, worries about illness, and hypochondriacal fears and beliefs. Psychotherapy and Psychosomatics, 61, 93–99. Stöber, J. (1998). Reliability and validity of two widely-used worry questionnaires: Selfreport and self–peer convergence. Personality and Individual Differences, 24, 887– 890. Stöber, J. (2000). Worry, thoughts, and images: A new conceptualization. In U. von Hecker, S. Dutke, & G. Sedak (Eds.), Generative mental processes and cognitive resources: Integrative research on adaptation and control (pp. 223–244). Dordrecht, The Netherlands: Kluwer. Street, L. L., Salman, E., Garfinkle, R., Silvestri, J., Carrasco, J., Cardenas, D., Zinbarg, R., Barlow, D. H., & Liebowitz, M. R. (1997). Discriminating between generalized anxiety disorder and anxiety disorder not otherwise specified in a Hispanic population: Is it only a matter of worry? Depression and Anxiety, 5, 1–6. Tallis, F., Eysenck, M., & Mathews, A. (1991). Elevated evidence requirements and worry. Personality and Individual Differences, 12, 21–27. Turk, C. L., Heimberg, R. G., Luterek, J. A., Mennin, D. S., & Fresco, D. M. (in press). Delineating emotion regulation deficits in generalized anxiety disorder: A comparison with social anxiety disorder. Cognitive Therapy and Research. Turner, S. M., Beidel, D. C., Borden, J. W., Stanley, M. A., & Jacob, R. G. (1991). Social phobia: Axis I and II correlates. Journal of Abnormal Psychology, 100, 102–106. Versiani, M., Mundim, F. D., Nardi, A. E., & Liebowitz, M. R. (1988). Tranylcypromine in social phobia. Journal of Clinical Psychopharmacology, 8, 279–283. Wells, A., & Carter, K. (2001). Further tests of a cognitive model of generalized anxiety

28

GAD IN THE CLINIC AND IN THE COMMUNITY

disorder: Metacognitions and worry in GAD, panic disorder, social phobia, depression, and nonpatients. Behavior Therapy, 32, 85–102. Wells, A., & Morrison, A. P. (1994). Qualitative dimensions of normal worry and normal obsessions: A comparative study. Behaviour Research and Therapy, 32, 867– 870. Wisocki, P. A. (1988). Worry as a phenomenon relevant to the elderly. Behavior Therapy, 19, 369–379. World Health Organization. (1992). International classification of diseases and related health problems (10th rev.). Geneva: Author. York, D., Borkovec, T. D., Vasey, M. W., & Stern, R. (1987). Effects of worry and somatic anxiety induction on thoughts, emotion and physiological activity. Behaviour Research and Therapy, 25, 523–526. Zebb, B. J., & Beck, J. G. (1998). Worry versus anxiety: Is there really a difference? Behavior Modification, 22, 45–61. Zimmerman, M., & Chelminski, I. (2003). Generalized anxiety disorder in patients wtih major depression: Is DSM-IV’s hierarchy correct? American Journal of Psychiatry, 160, 504–512.

GAD IN THE CLINIC AND IN THE COMMUNITY Epidemiology

CHAPTER 2

Epidemiology RONALD C. KESSLER ELLEN E. WALTERS HANS-ULRICH WITTCHEN

Although the notion of persistent pathological anxiety has been a central concept in psychiatry since the founding of the discipline in the 19th century, the American Psychiatric Association (APA) did not introduce the diagnosis of generalized anxiety disorder (GAD) in its Diagnostic and Statistical Manual of Mental Disorders (DSM) until slightly over two decades ago, with the publication of the third edition (DSM-III; APA, 1980). Prior to that time, GAD was conceptualized as one of the two main components of anxiety neurosis, the other component being panic (APA, 1968). The recognition that these two syndromes, though often occurring together, are sufficiently independent empirically to be considered distinct disorders caused the separation of panic disorder and GAD in DSM-III. However, DSM-III conceptualized GAD as the residual diagnosis in the pair and saw the clinical significance of GAD as fairly minor. The DSM-III definition of GAD required uncontrollable and diffuse (i.e., not focused on a single major life problem) anxiety or worry that was excessive or unrealistic in relation to objective life circumstances, and that persisted for 1 month or longer. A number of psychophysiological symptoms were required to occur with the anxiety or worry. Early clinical studies evaluating DSM-III GAD in clinical samples found that the disorder seldom occurred in the absence of some other comorbid anxiety or mood disorder (Breier, Charney, & Heninger, 1985; Breslau, 1985). This led some commentators to suggest that GAD might better be conceptualized as a prodrome, residue, or severity marker than as an independent disor29

30

GAD IN THE CLINIC AND IN THE COMMUNITY

der (Breslau & Davis, 1985b; Clayton et al., 1991; Noyes et al., 1992). Comorbidity between GAD and major depression was found to be lower as the duration of GAD increased (Breslau & Davis, 1985a), leading the DSM-III-R committee on GAD to require 6 months of symptoms for a diagnosis, rather than the 1 month required in DSM-III (APA, 1987). Further changes in the definition of excessive worry and the required number of associated symptoms were made in DSM-IV (APA, 1994).

PREVALENCE Changing definitions of GAD have made it difficult to collect long-term data on the prevalence and course of the disorder. Nonetheless, a critical mass of data has become available over the past decade showing that in the United States the general population lifetime prevalence of DSM-III and DSM-III-R GAD is between 4% and 7%; the 1-year prevalence is between 3% and 5%; and the current prevalence is between 1.5% and 3% (Blazer, Hughes, George, Swartz, & Boyer, 1991; Kessler, Dupont, Berglund, & Wittchen, 1999; Wittchen, Zhao, Kessler, & Eaton, 1994). Community epidemiological surveys of DSM-III-R GAD carried out in other parts of the world have found similar prevalences (e.g., Bijl, Ravelli, & Van Zessen, 1998; Offord et al., 1996). These prevalence estimates are considerably lower than the estimated prevalences of major depression, social phobia, or specific phobia, but are quite comparable to those of panic disorder, agoraphobia without panic disorder, and dysthymia. It should be noted that the prevalence estimates given above are based on DSM criteria. Whereas the DSM-III and DSM-III-R definitions of GAD required the worry to be excessive or unrealistic, and the DSM-IV definition requires the worry to be excessive, no such requirement exists in the International Classification of Diseases and Related Health Problems, 10th revision (ICD-10). One would consequently expect that the prevalence of GAD would be higher, possibly considerably higher, when ICD-10 criteria are used than when DSM-IV criteria are used. However, we are aware of no general population epidemiological research that has investigated this issue. General population surveys that evaluate the core dimensions of personality consistently find that substantial proportions of the population report anxious personality characteristics, such as persistent tension, worry, nervousness, and restlessness, in assessments of neuroticism (Heath, Neale, Kessler, Eaves, & Kendler, 1992; Zuckerman, 1991). The rates of GAD found in psychiatric epidemiological surveys, however, are much lower. This suggests that the DSM definitions of GAD exclude a substantial proportion of people with clinically significant chronic anxiety (i.e., neuroticism). The reason for this exclusion is unclear, but one possi-

Epidemiology

31

bility is that many people who score high on personality scales of neuroticism have brief recurrent episodes of anxiety that do not meet the requirement of occurring more days than not for 6 months. Another possibility is that the anxiety of people with high neuroticism scores is not associated with the full range of psychophysiological symptoms required for a DSM diagnosis of GAD. A recent report from the World Health Organization (WHO) study on Psychological Problems in Primary Care (Ustun & Sartorius, 1995) investigated these possibilities (Maier et al., 2000). The authors of this report attempted to validate the ICD-10 (WHO, 1990) requirements of a 6month duration and four associated symptoms for a diagnosis of GAD. Their attempt compared levels of role impairment associated with GADlike syndromes among “pure” cases (i.e., cases without other comorbid mental disorders) as a function of duration and number of associated symptoms. They found no evidence that impairment increased more after a duration of 6 months than at any other arbitrarily specified duration. In fact, revising the duration requirement to 1 month and lowering the number of required associated symptoms yielded the best discrimination between cases and noncases in terms of impaired role functioning. Because this investigation examined a primary care sample, it is unknown how much the estimated population prevalence of GAD would increase if these changes in diagnostic criteria were evaluated in a general population sample. The increase in prevalence would probably be substantial. From these results, it must be concluded that uncertainty exists about the true population prevalence of the GAD syndrome. Taking available studies of neurotic personality traits as a guide, the true current prevalence of the clinically significant component of the GAD syndrome could be as high as 5–8% of the general population, compared to the 1.5– 3.0% found in epidemiological studies of DSM GAD.

ONSET AND COURSE Retrospective reports from respondents in community epidemiological surveys (e.g., Burke, Burke, Rae, & Regier, 1991; Kendler, Neale, Kessler, Heath, & Eaves, 1992a) and from patients in clinical studies (e.g., Barlow, Blanchard, Vermilyea, Vermilyea, & DiNardo, 1986; Rogers et al., 1999) consistently suggest that GAD typically begins in the decade between the late teens and the late 20s. Only a small number of first onsets occur subsequent to the mid-30s. Examination of comparative Kaplan–Meier age-atonset curves in epidemiological samples suggests that specific phobia and social phobia are the only two anxiety or mood disorders with earlier ageat-onset distributions than GAD (Burke et al., 1991). It is noteworthy that childhood-onset cases of GAD are classified in some research as overanx-

32

GAD IN THE CLINIC AND IN THE COMMUNITY

ious disorder of childhood rather than GAD, leading to a reduction in the estimated prevalence of GAD and an increase in the estimated age at onset. Retrospective reports from respondents in community epidemiological surveys (e.g., Angst & Vollrath, 1991; Blazer et al., 1991) and from patients in clinical studies (Mancuso, Townsend, & Mercante, 1993; Noyes, Holt, & Woodman, 1996) also suggest that GAD is a very chronic condition, with episodes commonly persisting for a decade or longer. Prospective data show the same. The richest prospective data of this sort come from the Harvard/Brown Anxiety Research Program (HARP) (Warshaw, Keller, & Stout, 1994; Yonkers, Massion, Warshaw, & Keller, 1996). HARP is a prospective naturalistic study of patients with anxiety who were recruited initially from psychiatric clinics and hospitals in the Boston metropolitan area. In addition to obtaining detailed retrospective history data from these patients at baseline, HARP has thus far reported results for patients followed for 5 years (at 6-month intervals for the first 2 years and annually thereafter) to observe the natural history of their disorders. Only 15% of respondents with baseline GAD had a full remission for 2 months or longer at any time in the first year after baseline. This proportion increased to only 25% in the 2 years after baseline (Yonkers et al., 1996) and to 38% after 5 years (Yonkers, Dyck, Warshaw, & Keller, 2000). One would expect the course of GAD to be more persistent among patients in treatment than in the general population. It is noteworthy, therefore, that the community epidemiological data collected in the Epidemiologic Catchment Area (ECA) study (Blazer et al., 1991) are consistent with the results of the HARP study. Both show that the duration of DSM-III GAD is typically chronic. An indirect confirmation of this same conclusion can be obtained by examining the ratio of current prevalence to lifetime prevalence in community epidemiological surveys. These ratios are typically in the range between 40% and 60% (e.g., Bijl et al., 1998; Offord et al., 1994). Although these ratios cannot distinguish the effects of chronicity from recurrence, such high ratios clearly imply that people with a history of GAD spend a substantial proportion of their subsequent lifetimes with the disorder.

SOCIODEMOGRAPHIC CORRELATES OF ONSET AND COURSE Epidemiological surveys show that the lifetime prevalence of GAD is more common among women than men, among unmarried than married people, among racial/ethnic minorities than members of majority groups, and among respondents with low socioeconomic status (SES) than middle- or high-SES respondents (Blazer et al., 1991; BrawmanMintzer & Lydiard, 1996; Wittchen et al., 1994). However, prospective analyses show that none of these sociodemographic variables is a signifi-

Epidemiology

33

cant predictor of the course of GAD (Yonkers et al., 2000). Consistent with this finding, unpublished analyses of cross-sectional epidemiological data from the WHO’s International Consortium in Psychiatric Epidemiology (ICPE, 2000) show that none of these sociodemographic predictors is significantly related to the current prevalence of GAD among respondents with a lifetime history of the disorder, after age at onset and time since onset are controlled for. In the case of acquired statuses like SES and marital status, where significant associations exist with lifetime prevalence, the absence of significant associations with course could be due to selection processes; that is, GAD or its determinants might inf luence subsequent SES or marital status. For example, whereas low SES may predispose an individual to excessive worry, excessive worry in some cases may lead to increased diligence that promotes increased education and occupational achievement, leading to a canceling of an association between SES and course of GAD. In the case of ascribed statuses that do not change over time, such as gender and race/ethnicity, a more plausible interpretation is that these statuses or their correlates do not affect the course of GAD, though they are etiologically significant for onset. A complex nonlinear cross-sectional association also exists between GAD and age, with the highest current prevalences in the middle age groups (Wittchen et al., 1994). Decomposition shows that this association is caused by the joint occurrence of significant cohort effects (i.e., increasing prevalences among people born in more recent years, presumably due to the increase in environmental risk factors) and differential recurrence risk as a function of age (i.e., increasing risk of GAD’s becoming persistent, related to increasing age). The cohort effect is inferred from comparisons of retrospective age-at-onset reports in cross-sectional community epidemiological surveys. These analyses suggest that the lifetime prevalence of GAD has been on the rise in successive cohorts born over the past half century. At least part of this presumed increase may be due to differential recall failure with age. However, the fact that the strength of the retrospectively reported intercohort difference in GAD is greater than for other disorders of comparable impairment argues against recall bias as a complete explanation. Once onset occurs, risk of current prevalence is inversely related to age but not to age at onset. Consequently, the highest current prevalence is in the middle years of life, due to a comparatively high lifetime risk (cohort effect) and a comparatively low rate of remission. Perhaps the most intriguing of the sociodemographic correlates of GAD involves SES. This association has special interest because of preliminary evidence from the ICPE surveys that the association is stronger when ICD criteria are used instead of DSM criteria. The reason for this is that ICD, unlike DSM, does not require that persistent worry be excessive in order to qualify for a diagnosis of GAD. As one might expect, socioeco-

34

GAD IN THE CLINIC AND IN THE COMMUNITY

nomically disadvantaged people are much more likely than other people to experience chronic life situations that lead to clinically significant persistent anxiety (e.g., housing instability or neighborhood violence). Because these anxieties are often justified (i.e., not excessive in light of objective life circumstances), DSM criteria rule out a diagnosis of GAD, dampening the association between SES and GAD.

COMORBIDITY As noted at the beginning of the chapter, early clinical studies documented very high rates of comorbidity among patients with DSM-III GAD. This led to an expansion of the duration requirement from 1 month to 6 months in subsequent editions of the DSM. However, clinical studies using DSM-III-R criteria found that comorbidity remained quite high even with the 6-month duration requirement (e.g., Noyes et al., 1992; Roy-Byrne, 1996). This led to continued suggestions that GAD is not an independent disorder (Brawman-Mintzer et al., 1993; Gorman, 1996). It is important to appreciate, though, that this evidence of high comorbidity comes almost entirely from clinical samples. Studies of GAD comorbidity carried out in community samples suggest that the initial concern about extremely high comorbidity in GAD was misplaced. Wittchen et al. (1994) first documented this in an analysis of a large community epidemiological survey, which showed that the proportion of people with DSM-III-R GAD and one or more other comorbid diagnoses was not dramatically higher than the rate of comorbidity among people with other anxiety or mood disorders. Wittchen and his colleagues explained the discrepancy between this result and earlier clinical findings by showing that comorbidity was a powerful predictor of help seeking among people with GAD. This suggests that the extremely high comorbidity in GAD is an artifact of treatment sample selection bias rather than an inherent feature of GAD. A similar result regarding comorbidity in GAD is found by looking at lifetime comorbidity rather than episode comorbidity. Table 2.1 illustrates this result by presenting data from the U.S. National Comorbidity Survey (NCS) (Kessler et al., 1994), a large nationally representative survey of the U.S. household population. Table 2.1 shows the proportions of NCS respondents with particular lifetime DSM-III-R disorders who also met criteria for at least one other lifetime DSM-III-R comorbid disorder. As shown in the table, proportional lifetime comorbidity in GAD was not meaningfully higher than in dysthymia or panic disorder and was considerably lower than in mania. Before we assume that this finding shows that comorbidity in GAD is no higher than in other anxiety or mood disorders, it is important to consider comparative comorbidities in terms of bivariate associations (e.g., sim-

Epidemiology

35

TABLE 2.1. Proportions of National Comorbidity Survey (NCS) DSM-III-R Disorders with Lifetime Comorbidity Mood disorders Major depressive episode Dysthymia Mania Any mood disorder

83.1% 91.3% 99.4% 82.2%

Anxiety disorders Generalized anxiety disorder Panic disorder Social phobia Simple phobia Agoraphobia Posttraumatic stress disorder Any anxiety disorder

91.3% 92.2% 81.0% 83.4% 87.3% 81.0% 74.1%

Note. Disorders are operationalized without diagnostic hierarchy rules. Reprinted from Table 2-1 in “The Prevalence of Psychiatric Comorbidity” (p. 28) by R. C. Kessler, 1997, in S. Wetzler & W. C. Sanderson (Eds.), Treatment Strategies for Patients with Psychiatric Comorbidity. Copyright 1997 by John Wiley & Sons, Inc. This material was used by permission of John Wiley & Sons, Inc.

ple Pearson correlations, tetrachoric correlations, or odds ratios [ORs]) rather than in terms of conditional probabilities (i.e., conditional risks of one disorder, given the existence of another disorder). This is important, because conditional probabilities are inf luenced by differential prevalences. Analyzing the data in Table 2.1 this way reveals bivariate lifetime and 12month comorbidities, expressed as ORs for GAD (with an average OR of 12.0) that are no larger than those for major depression (with an average OR of 12.5) or panic disorder (with an average OR of 12.0) (Kessler, 1995). An OR of 12 means that the odds of having versus not having one of the disorders in the pair are 12 times as high among people who have the other disorder than among those who do not have the other disorder.

SYMPTOM SPECIFICITY An issue of special importance in the evaluation of comorbidity is whether the symptoms of a presumably distinct disorder form separate empirical clusters from symptoms of other disorders in representative samples (Robins & Guze, 1970). Maier and his colleagues (2000) evaluated this issue in the WHO study on Psychological Problems in Primary Care. They found that the associated psychophysiological symptoms of GAD specified in ICD-10 strongly clustered with the core symptoms of persistent worry and anxiety.

36

GAD IN THE CLINIC AND IN THE COMMUNITY

Comorbidity between GAD and major depression has been the focus of special interest, based on evidence that this is the most common type of anxiety–mood comorbidity (Kessler, 1997). One reason for this high comorbidity is presumably that the two disorders share a number of common symptoms. Therefore, it is noteworthy that investigations of comorbidity carried out at the symptom level rather than at the diagnostic level are able to find differentiation of GAD and major depression. Brown, Chorpita, and Barlow (1998) tested several models of structural relationships among symptoms of GAD and major depression, and found separate latent factors of positive affectivity, negative affectivity, and autonomic suppression (related to GAD). This finding strongly argues that GAD and major depression can be distinguished, despite the overlap of some core symptoms. Consistent with this finding, analyses of twin data using an additive behavior-genetic model conclude that the environmental determinants of GAD and major depression are distinct (Kendler, Neale, Kessler, Heath, & Eaves, 1992b). In turn, this result is consistent with the finding in epidemiological research that GAD and major depression have significantly different sociodemographic predictors (Skodal, Schwartz, Dohrenwend, Levav, & Shrout, 1994). For example, SES is much more strongly related to major depression than to GAD. It is noteworthy that twin studies also suggest that the genes for GAD and major depression are the same (Kendler et al., 1992b), raising the possibility that the two syndromes are different manifestations of the same underlying disorder. However, the model on which this conclusion is based assumes that the joint effects of genes and environment are additive—in other words, that the impact of environmental determinants is not inf luenced by the presence or absence of genes. This is implausible. A more realistic interactive specification, which cannot be identified with conventional twin data, might well show differentiation of genetic effects. Consistent with this possibility, studies show differential aggregation of mental disorders in the families of patients with GAD and major depression (Reich, 1993); this finding suggests strongly that GAD and major depression are distinct disorders (Reich, 1995).

TEMPORAL PRIORITIES IN LIFETIME COMORBID GAD Another important issue in evaluating comorbid conditions concerns the distinction between primary and secondary disorders. Although the primary–secondary distinction is defined in a number of ways, one of the most simple and intuitive of these definitions compares ages of onset. Based on NCS data, Table 2.2 shows the proportion of lifetime cases of specific DSM-III-R disorders that were temporally primary—that is, the proportion in which the disorder in question was the earliest lifetime dis-

Epidemiology

37

TABLE 2.2. Proportions of Temporally Primary Cases among NCS DSM-III-R Disorders Mood disorders Major depressive episode Dysthymia Mania

41.1% 37.7% 20.2%

Anxiety disorders Generalized anxiety disorder Panic disorder Social phobia Simple phobia Agoraphobia Posttraumatic stress disorder

37.0% 23.3% 63.1% 67.6% 45.2% 52.1%

Note. Disorders are operationalized without diagnostic hierarchy rules. Reprinted from Table 2-3 in “The Prevalence of Psychiatric Comorbidity” (p. 32) by R. C. Kessler, 1997, in S. Wetzler & W. C. Sanderson (Eds.), Treatment Strategies for Patients with Psychiatric Comorbidity. Copyright 1997 by John Wiley & Sons, Inc. This material was used by permission of John Wiley & Sons, Inc.

order ever experienced by the respondent, according to retrospective ageat-onset reports. In cases where two or more disorders reportedly started at the same age and before any other disorders, both were coded as temporally primary. The results are surprising in showing that simple phobia, social phobia, and posttraumatic stress disorder were the only DSM-III-R anxiety or mood disorders for which the majority of lifetime cases were temporally primary. In general, anxiety disorders are more likely than mood disorders to be temporally primary. The proportion temporally primary for GAD (37.0%) was comparable to the percentages for the mood disorders (20.2–41.1%) and slightly higher than for panic disorder (23.3%). Although not reported here, more detailed analyses of these data have shown that GAD was typically the temporally primary disorder in relation to depression among people with comorbid anxious–depression (Kessler et al., 1996).

PREDICTIVE PRIORITIES IN LIFETIME COMORBID GAD It is a mistake to think of temporal priority as equivalent to causal priority. The easiest way to make this clear is to note that it is possible to have a pair of comorbid disorders in which the prior occurrence of Disorder A is not a significant predictor of the subsequent onset of Disorder B, even

38

GAD IN THE CLINIC AND IN THE COMMUNITY

though Disorder A occurs before Disorder B in the vast majority of cases (Kessler & Price, 1993). This situation can occur when the base rates of the two disorders differ substantially, as with comorbidity between GAD and major depression (where major depression is much more common than GAD). A more useful way to think about priorities in comorbidity is to investigate prediction: whether one disorder in a particular pair, when it is temporally primary, is a significant predictor of the subsequent first onset of the other disorder. The NCS examined this by using retrospective age-atonset reports for each disorder to estimate a series of bivariate survival models, in which prior onset of one disorder was treated as a time-varying covariate of another disorder (Kessler, 1997). For ease of comparison, these models constrained the effects of temporally primary disorders to be constant across ages of onset, times since onset, and age at onset of the earlier disorders. All temporally primary anxiety and mood disorders were statistically significant predictors of the subsequent first onset of other anxiety and mood disorders. GAD did not stand out in any particular way in comparison to the other disorders—either in the magnitude of the effects of GAD in predicting later onset of other disorders, or in the magnitude of the effects of other temporally primary disorders in predicting first onset of GAD. This means that any questioning of the validity of GAD as a diagnostic entity on the basis of data about predictive priority in comorbidity would apply equally to major depression, panic disorder, social phobia, and all the other anxiety and mood disorders considered in this analysis.

PREDICTIVE PRIORITIES INVOLVING SEVERITY AND COURSE NCS analyses also investigated whether comorbidity is associated more strongly with the severity or course of GAD than of other anxiety or mood disorders (Kessler, 2000). The rationale was that if GAD is a prodrome, residue, or severity marker of other disorders, the severity and course of GAD should be much more strongly affected by comorbidity than those of other anxiety or mood disorders. The results showed that comorbidity was generally associated with increased severity and persistence for all anxiety and mood disorders. Although the patterns involving GAD were generally similar to those for the other disorders considered in the analysis, there was one important exception. GAD was the only anxiety or mood disorder in which persistence, as indirectly indicated by recency (with age at onset and time since onset controlled for), was unrelated to comorbidity. Yonkers and colleagues (2000) reported a similar result in an analysis of the prospective study of the predictors of the clinical course of GAD in

Epidemiology

39

the HARP study. The HARP data also showed that the course of comorbid GAD was unrelated to whether the GAD was primary or secondary (Rogers et al., 1999). Other studies arrived at different conclusions about the effects of comorbidity on the course of GAD (Angst & Vollrath, 1991; Durham, Allan, & Hackett, 1997; Mancuso et al., 1993). However, the inconsistency of this evidence—in conjunction with the clear and consistent evidence that comorbidity is significantly related to the course of other anxiety and mood disorders—means that, if anything, GAD behaves more like an independent disorder in this respect than do other anxiety or mood disorders. This finding is relevant to the hierarchy requirement in DSM-IV that periods of generalized anxiety occurring exclusively within episodes of major depression do not qualify for a diagnosis of GAD. It is noteworthy that there is no symmetrical hierarchy rule in DSM-IV that periods of depression occurring exclusively within episodes of generalized anxiety do not qualify for a diagnosis of major depression. Given that GAD is usually temporally primary to major depression, and that episodes of GAD generally persist much longer than episodes of major depression, the number of people with GAD whose anxiety occurs exclusively within episodes of major depression is quite small. However, a sizable number of people with major depression may have episodes of depression that occur exclusively within longer episodes of anxiety. It would be useful for future research to investigate subtypes of depression along this dimension.

THE IMPAIRMENTS ASSOCIATED WITH PURE AND COMORBID GAD As previously noted, Wittchen and his associates found the false appearance of high comorbidity in GAD in early clinical studies to be due to an exceptionally strong help-seeking bias among people with comorbid GAD. People in the general population with GAD had extremely low rates of help seeking in the absence of some other comorbid mental disorder. There are at least two plausible interpretations of this finding. One is that pure GAD is not seriously impairing in itself; only when GAD co-occurs with other anxiety or mood disorders does the level of distress in GAD motivate people with the disorder to seek treatment. If this is the case, it could be argued that GAD may be an independent disorder, but it is only clinically significant when it is part of a comorbid cluster in which the comorbid disorders should be the focus of clinical attention. The other plausible interpretation is that people with GAD worry about a great many things, but seldom perceive that their worrying is a problem without the presence of other comorbid disorders that are more easily recognized as pathological. The distinction between these two possibilities is of con-

40

GAD IN THE CLINIC AND IN THE COMMUNITY

siderable importance in evaluating whether GAD should be considered a clinically significant disorder in its own right. Three recent studies focused on pure and comorbid GAD in primary care samples and obtained results relevant to this issue (Olfson et al., 1997; Ormel et al., 1994; Schonfeld et al., 1997). All three studies found that “pure” GAD, defined as a current episode of this disorder in the absence of any of the other mood, anxiety, or substance use disorders assessed in the primary care samples, was associated with meaningful levels of impairment in a number of life domains. In the largest of these studies, Ormel and colleagues (1994) found that mean numbers of disability days in the past month were much higher among primary care patients with pure GAD (4.4) than among patients with none of the psychiatric disorders assessed in their survey (1.7). Schonfeld and colleagues (1997) found that mean age- and sex-adjusted scores on the 0–100 Short Form-36 scale of social functioning (Stewart, Hays, & Ware, 1988) were much lower (with a high score indicating good functioning) among primary care patients with pure GAD (71.0) than among those with none of the psychiatric disorders assessed in their survey (83.6). Given the importance placed on major depression in previous studies of comorbid GAD, it is noteworthy that the impairment associated with pure major depression was greater and more varied than that associated with pure GAD in two of these studies (Olfson et al., 1997; Schonfeld et al., 1997). However, the reliability of these findings can be called into question, based on the fact that the numbers of patients with pure disorders in these studies were quite small: only between 4 and 14 cases of pure GAD, and only between 25 and 54 cases of pure depression. Differences in the magnitude of impairment associated with pure GAD versus pure depression were much smaller in the study carried out by Ormel and colleagues (1994), which included considerably larger numbers of respondents with pure GAD (n = 272) and pure depression (n = 438). Aggregate impairments of GAD and major depression were also found to be quite similar to each other in a large primary care study conducted by Spitzer and colleagues (1995) that did not distinguish between pure and comorbid cases. It is important to note that the evaluations of comorbidity in all these primary care studies included other psychiatric disorders in addition to GAD and depression. Given the higher prevalence of comorbidity between GAD and depression than of other types of anxiety–mood comorbidity (Kessler, 1997), more focused general population data are necessary to evaluate the separate and joint effects of these two disorders, as well as to evaluate the effects of broader comorbidities in the general population. This was recently done in a comparative analysis of pure and comorbid GAD and depression in two large nationally representative epidemiological surveys in the United States (Kessler et al., 1999).

Epidemiology

41

Table 2.3 presents data reproduced from the investigation on the associations between 12-month GAD without depression and 12-month major depression without GAD and impairment, adjusted for other comorbid disorders and for sociodemographic variables. The first two columns report associations involving GAD without depression. All six of these coefficients show that GAD was associated with higher impairments than found among respondents who did not have GAD (three of the six were statistically significant). The next two columns report associations involving major depression without GAD. All of these coefficients show that depression was associated with significantly higher impairment than found among respondents who did not have depression. Finally, comparison of the impairment associated with GAD without depression versus depression without GAD is reported in the last two columns of the table. None of these differences was statistically significant. Data from the same report found that anxiety–depression was associated with higher impairment than found among respondents with only one of the two disorders. This report also noted that the magnitude of this impairment did not differ depending on which of the two disorders was temporally primary, either on a lifetime basis or in terms of episode onset.

TABLE 2.3. The Effects (Odds Ratios) of 12 Month GAD without Major Depression (MD) and MD without GAD in Predicting Impairments in Two U.S. National Surveys,a Controlling for Sociodemographics and Other 12-Month DSM-III-R Disordersb

GAD without MD Survey Survey 1 2 Fair/poor perceived mental health High work impairment High social impairment

6.0* 3.5 2.5*

4.8* 3.5 1.2

MD without GAD Survey Survey 1 2 3.3* 3.5* 2.0*

5.2* 8.5* 1.6*

GAD without MD vs. MD without GAD Survey Survey 1 2 1.6 0.9 1.5

0.8 0.5 1.0

Note. Reprinted from Table 2 in “Impairment in Pure and Comorbid Generalized Anxiety Disorder and Major Depression at 12 Months in Two National Surveys” by R. C. Kessler, R. L. DuPont, P. Berglund, & H. U. Wittchen, 1999, American Journal of Psychiatry, 156, p. 1919. Copyright 1999, the American Psychiatric Association. Reprinted by permission. a The two surveys indicated here are the NCS (Survey 1; Kessler et al., 1994) and the Midlife Development in the U.S. Survey (Survey 2; Kessler et al., 1999). b Results are based on separate regression equations evaluating the effect of either GAD or major depression in predicting one of the impairment measures in one of the samples while controlling for sociodemographic variables (age, gender, education, race/ethnicity, employment status, marital status, and urbanicity) and other 12-month DSM-III-R disorders. Models in the first two columns evaluate the effect of 12-month GAD on the subsample of respondents who did not have 12-month MD. Models in the middle two columns evaluate the effect of 12-month MD on the subsample of respondents who did not have 12-month GAD. Models in the last two columns evaluate the relative impairments of GAD without MD versus MD without GAD in analyses that are confined to respondents in those two subsamples. * Significant at the .05 level (two-sided test).

42

GAD IN THE CLINIC AND IN THE COMMUNITY

PATTERNS OF TREATMENT SEEKING As noted earlier in the chapter, the NCS data showed clearly that comorbidity was a powerful predictor of help seeking among people with GAD (Wittchen et al., 1994). It is important to appreciate that this finding was not obtained because pure GAD is not severe enough to motivate help-seeking efforts. The results presented in the preceding section have shown that pure GAD is as impairing as pure major depression. Why, then, do a much higher proportion of people with pure depression than with pure GAD seek professional help? One plausible explanation for this difference is that GAD often starts so early in life and has such an insidious onset, often associated with underlying personality pathology (Blashfield et al., 1994), that many people with GAD do not realize that they have an emotional problem until they develop secondary disorders. Olfson, Kessler, Berglund, and Lin (1998) contributed one piece of evidence that is indirectly consistent with this interpretation. They documented that speed of initial treatment contact after first onset of GAD was inversely related to age at onset of the disorder in general population samples in both the United States and Ontario. Figure 2.1 presents the aggregate speed of contact curves for GAD reported by Olfson and colleagues, while Table 2.4 presents the effects of age at onset. As shown in Figure 2.1, approximately one-third of people with GAD sought treatment in the year of onset of the disorder, whereas the average delay in initial treatment contact was more than a decade among people who delayed beyond the first year. This was true not only in the United States, where there are financial barriers to seeking mental health treatment, but also in Ontario, where such barriers are substantially lower. This result argues against the notion that financial barriers to treatment are an important determinant of treatment delays. The finding that age at onset was associated with speed of help seeking, as shown in Table 2.4, can most plausibly be atrributed to an increased recognition of anxiety as a problem with increasing age at onset. The especially strong age-at-onset effect for seeking treatment in the year of onset is consistent with the finding of HoehnSaric, Hazlett, and McLeod (1993) that the vast majority of people with early-onset GAD reported an insidious onset, while people with lateronset GAD were more likely to report an acute onset associated with a precipitating stressful event. These pervasive delays in help seeking are especially upsetting, in light of the fact that effective cognitive-behavioral (Barlow, 2002) and pharmacological (Hackett, 2000) therapies exist for the treatment of GAD. There is also promising evidence regarding the efficacy of psychodynamic treatments for GAD (Crits-Christoph, Connolly, Azarian, Crits-Christoph, & Shappell, 1996). Perhaps even more upsetting is the fact that only a small proportion of people with GAD get treatment, even

Epidemiology

43

100% 90%

Cumulative Probability of Treatment

Ontario (N=153) 80%

United States (N=368)

70% 60% 50% 40% 30% 20% 10% 0% 05

10

15

20

25

30

35

40

45

Years Since Onset of Disorder

FIGURE 2.1. Speed-of-contact curves for the probability of initial treatment contact after first onset of GAD in the United States and Ontario. Reprinted from Part 2 of Figure 1 in “Psychiatric Disorder Onset and First Treatment Contact in the United States and Ontario” by M. Olfson, R. C. Kessler, P. A. Berglund, & E. Lin, 1998, American Journal of Psychiatry, 155, p. 1416. Copyright 1998, the American Psychiatric Association. Reprinted by permission.

though the majority of them come into frequent contact with the primary care system. Indeed, GAD is the most commonly found anxiety disorder in primary care settings, both in the United States (Barrett, Oxman, & Gerber, 1988) and in many other parts of the world (Ustun & Sartorius, 1995). This is due to the fact that people with GAD frequently present for treatment with complaints of minor physical problems. The problem is that their anxiety is seldom taken as a focus of treatment, because primary care patients with GAD seldom include anxiety among their presenting complaints, and primary care physicians seldom recognize anxiety in their patients unless it is a presenting complaint (Ormel et al., 1990). This lack of recognition of the underlying psychological problem leads to frustrations on the part of patients, who typically are not relieved of the secondary psychophysiological symptoms that are usually their presenting complaints (Katon et al., 1990; Lin et al., 1991). Patient screening and demand management initiatives are currently being evaluated in an effort to increase recognition and treatment of GAD among these patients. Another source of considerable concern regarding treatment of GAD is that research shows that a substantial proportion of patients in treat-

44

GAD IN THE CLINIC AND IN THE COMMUNITY TABLE 2.4. The Effects of Age at Onset on Speed of First Treatment Contact for GAD in the United States and Ontario Odds ratiosa for treatment contact In the year of onset of GAD In subsequent years Age at onset £ 12 13–19 20–29 ³ 30

U.S.

Ontario

U.S.

Ontario

1.0 22.5 73.6 92.5

1.0 7.5 35.8* 21.2*

1.0 1.4 2.0 5.6*

1.0 1.3 2.0 2.3

Note. Adapted from Tables 1 and 2 in “Psychiatric Disorder Onset and First Treatment Contact in the United States and Ontario” by M. Olfson, R. C. Kessler, P. A. Berglund, & E. Lin, 1998, American Journal of Psychiatry, 155, pp. 1417–1418. Copyright 1998, the American Psychiatric Association. Adapted by permission. a These odds ratios were obtained by exponentiating logistic regression coefficients (in the model to predict contact in the year of onset of GAD) and by discrete-time survival coefficients (in the model to predict contact in subsequent years) from a model that included controls for cohort (in both) and time since onset (in the model to predict contact in subsequent years). There were 368 cases of GAD in the U.S. sample and 153 in the Ontario sample. * Significant at the .05 level (two-sided test).

ment receive inadequate treatment. In a follow-up of the HARP sample, Goisman, Warshaw, and Keller (1999) showed that as of 1996, dynamic psychotherapy remained the most commonly used form of talk therapy for patients with GAD. They found that only a small number of patients received cognitive-behavioral therapies, even though treatment trials have documented the effectiveness of cognitive-behavioral therapies for GAD, whereas no such data exist yet for psychodynamic therapies. In a parallel analysis of the HARP data, Salzman, Goldenberg, Bruce, and Keller (2001) showed that up to one-third of patients seeking medical treatment for GAD in 1996 were unmedicated at the time of interview. An unspecified additional proportion were undermedicated. Consistent with these results, Wang, Berglund, and Kessler (2000), analyzed data from a large U.S. national telephone–mail survey conducted in 1996–1997 and found that the majority of patients in treatment for anxiety or mood disorders during those years received treatments that did not meet the minimum standards for effectiveness stipulated in published treatment guidelines.

OVERVIEW This chapter shows that GAD is a commonly occurring mental disorder that typically has an early age at onset, a chronic course, and a high degree of comorbidity with other anxiety and mood disorders. Comorbid

Epidemiology

45

GAD is often temporally primary, especially in relation to mood disorders. The weight of evidence reviewed here argues against the view that GAD is better conceptualized as a prodrome, residue, or severity marker of other disorders than as an independent disorder in its own right. Recent community epidemiological studies and primary care studies show consistently that comorbidity in GAD is no greater than in most other anxiety or mood disorders when comorbidities are evaluated in unrestricted samples. Furthermore, the data reviewed here show that pure GAD is associated with impairments comparable in magnitude to the impairments associated with other clinically significant anxiety and mood disorders. In addition, the clinical course of GAD is less consistently related to comorbidity than is the course of other anxiety and mood disorders. These findings argue that the status of GAD as an independent disorder is at least as strongly supported by available evidence as are other anxiety or mood disorders. The chapter has also shown that there is considerable uncertainty regarding even very basic epidemiological characteristics of the GAD syndrome. Lingering concerns about the independence of GAD have conspired to exacerbate this problem by promoting repeated changes in the diagnostic criteria for GAD from DSM-III to DSM-III-R and to DSM-IV. These successive changes have made it difficult to amass consistent longterm data on the natural history of GAD. Available evidence on diagnostic thresholds for GAD raises questions about whether the requirements of a 6-month minimum duration and the specific number of psychophysiological symptoms are optimal for capturing all of the people in the general population who suffer from a clinically significant generalized anxiety syndrome. An additional source of potential bias in this regard is the fact that the DSM system requires anxiety to be “excessive” to qualify as GAD. It is interesting to note that there is no comparable DSM requirement that dysphoria must be excessive to qualify as major depression. These uncertainties as to the most appropriate diagnostic threshold regarding duration and number of symptoms, and the discrepancy between the diagnostic requirements for anxiety and depression, make it very difficult to gain a clear understanding of the true breadth and depth of the GAD syndrome in the population. Future work is needed, ideally in unbiased epidemiological samples, to resolve these basic uncertainties. Given the strong comorbidity between GAD and major depression, the fact that the majority of people with this type of comorbidity report that the onset of GAD occurred prior to the onset of depression, and the fact that temporally primary GAD significantly predicts the subsequent onset of depression and other secondary disorders, a question can be raised as to whether early intervention and treatment of primary GAD might be effective in preventing the subsequent first onset of secondary anxiety and depression. Unfortunately, little is known about this possibility, due to the fact (noted above) that very few people with pure GAD seek

46

GAD IN THE CLINIC AND IN THE COMMUNITY

treatment. We do not understand why this is so. Given the early onset of GAD and its strong effects in predicting the subsequent onset, severity, and persistence of other disorders, efforts are needed to collect epidemiological data on the reasons for low help seeking in pure GAD, and to develop outreach strategies that might be able to correct this situation.

ACKNOWLEDGMENTS Preparation of this chapter was supported by U.S. Public Health Service Grant No. U01-MH60220 and by the Global Research on Anxiety and Depression Network (GRAD), through an unrestricted educational grant from Wyeth-Ayerst Pharmaceuticals (http://www.gradnetwork.com). The helpful comments of Sergio Aguilar-Gaxiola, Kathleen Merikangas, and Bedirhan Ustun are gratefully acknowledged.

REFERENCES American Psychiatric Association (APA). (1968). Diagnostic and statistical manual of mental disorders (2nd ed.). Washington, DC: Author. American Psychiatric Association (APA). (1980). Diagnostic and statistical manual of mental disorders (3rd ed.). Washington, DC: Author. American Psychiatric Association (APA). (1987). Diagnostic and statistical manual of mental disorders (3rd ed., rev.). Washington, DC: Author. American Psychiatric Association (APA). (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author. Angst, J., & Vollrath, M. (1991). The natural history of anxiety disorders. Acta Psychiatrica Scandinavica, 141, 446–452. Barlow, D. H. (2002). Anxiety and its disorders (2nd ed.). New York: Guilford Press. Barlow, D. H., Blanchard, R. B., Vermilyea, J. B., Vermilyea, B. B., & DiNardo, P. A. (1986). Generalized anxiety and generalized anxiety disorder: Description and reconceptualization. American Journal of Psychiatry, 143, 40–44. Barrett, J., Oxman, T. E., & Gerber, P. D. (1988). The prevalence of psychiatric disorders in primary care practice. Archives of General Psychiatry, 45, 1100–1106. Bijl, R. V., Ravelli, R., & Van Zessen, G. (1998). Prevalence of psychiatric disorder in the general population: Results of the Netherlands Mental Health Survey and Incidence Study (NEMESIS). Social Psychiatry and Psychiatric Epidemiology, 33, 587– 595. Blashfield, R., Noyes, R., Reich, J., Woodman, C., Cook, B. L., & Garvey, M. (1994). Personality disorder traits in generalized anxiety and panic disorder patients. Comprehensive Psychiatry, 35, 329–334. Blazer, D. G., Hughes, D., George, L. K., Swartz, M., & Boyer, R. (1991). Generalized anxiety disorder. In L. N. Robins & D. A. Regier (Eds.), Psychiatric disorders in America (pp. 180–203). New York: Free Press. Brawman-Mintzer, O., & Lydiard, R. B. (1996). Generalized anxiety disorder: Issues in epidemiology. Journal of Clinical Psychiatry, 57(Suppl. 7), 3–8. Brawman-Mintzer, O., Lydiard, R. B., Emmanuel, N., Payeur, R., Johnson, M., Rob-

Epidemiology

47

erts, J., Jarrell, M. P., & Ballenger, J. C. (1993). Psychiatric comorbidity in patients with generalized anxiety disorder. American Journal of Psychiatry, 150, 1216–1218. Breier, A., Charney, D. S., & Heninger, G. R. (1985). The diagnostic validity of anxiety disorders and their relationship to depressive illness. American Journal of Psychiatry, 142, 787–797. Breslau, N. (1985). Depressive symptoms, major depression and generalized anxiety: A comparison of self-reports on CES-D and results from diagnostic interviews. Psychiatry Research, 15, 219–229. Breslau, N., & Davis, G. C. (1985a). DSM-III generalized anxiety disorder: An empirical investigation of more stringent criteria. Psychiatry Research, 15, 231–238. Breslau, N., & Davis, G. C. (1985b). Further evidence on the doubtful validity of generalized anxiety disorder [Letter]. Psychiatry Research, 16, 177–179. Brown, T. A., Chorpita, B. F., & Barlow, D. H. (1998). Structural relationships among dimensions of the DSM-IV anxiety and mood disorders and dimensions of negative affect, positive affect, and autonomic arousal. Journal of Abnormal Psychology, 107, 179–192. Burke, K. C., Burke, J. D., Jr., Rae, D. S., & Regier, D. A. (1991). Comparing age at onset of major depression and other psychiatric disorders by birth cohorts in five U.S. community populations. Archives of General Psychiatry, 48(9), 789–795. Clayton, P. J., Grove, W. M., Coryell, W., Keller, M. B., Hirschfeld, R., & Fawcett, J. (1991). Follow-up and family study of anxious depression. American Journal of Psychiatry, 148, 1512–1517. Crits-Christoph, P., Connolly, M. B., Azarian, K., Crits-Christoph, K., & Shappell, S. (1996). An open trial of brief supportive–expressive psychotherapy in the treatment of generalized anxiety disorder. Psychotherapy, 33, 418–430. Durham, R. C., Allan, T., & Hackett, C. A. (1997). On predicting improvement and relapse in generalized anxiety disorder following psychotherapy. British Journal of Clinical Psychology, 36, 101–119. Goisman, R. M., Warshaw, M. G., & Keller, M. B. (1999). Psychosocial treatment prescriptions for generalized anxiety disorder, panic disorder, and social phobia, 1991–1996. American Journal of Psychiatry, 156, 1819–1821. Gorman, J. M. (1996). Comorbid depression and anxiety spectrum disorders. Depression and Anxiety, 4, 160–168. Hackett, D. (2000). Venlafaxine XR in the treatment of anxiety. Acta Psychiatrica Scandinavica, 406, 30–35. Heath, A. C., Neale, M. C., Kessler, R. C., Eaves, L. J., & Kendler, K. S. (1992). Evidence for genetic inf luences on personality from self-reports and informant ratings. Journal of Personality and Social Psychology, 63(1), 85–96. Hoehn-Saric, R., Hazlett, R. L., & McLeod, D. R. (1993). Generalized anxiety disorder with early and late onset of anxiety symptoms. Comprehensive Psychiatry, 34, 291– 298. International Consortium of Psychiatric Epidemiology (ICPE). (2000). Cross-national comparisons of the prevalences and correlates of mental disorders: Results from the WHO International Consortium in Psychiatric Epidemiology. Bulletin of the World Health Organization, 78, 413–426. Katon, W., von Korff, M., Lin, E., Lipscomb, P., Russo, J., Wagner, E., & Polk, E. (1990). Distressed high utilizers of medical care: DSM-III-R diagnoses and treatment needs. General Hospital Psychiatry, 12, 355–362. Kendler, K. S., Neale, M. C., Kessler, R. C., Heath, A. C., & Eaves, L. J. (1992a). Gener-

48

GAD IN THE CLINIC AND IN THE COMMUNITY

alized anxiety disorder in women: A population based twin study. Archives of General Psychiatry, 49, 267–272. Kendler, K. S., Neale, M. C., Kessler, R. C., Heath, A. C., & Eaves, L. J. (1992b). Major depression and generalized anxiety disorder: Same genes, (partly) different environments? Archives of General Psychiatry, 49, 716–722. Kessler, R. C. (1995). The epidemiology of psychiatric comorbidity. In M. Tsuang, M. Tohen, & G. E. P. Zahner (Eds.), Textbook of psychiatric epidemiology (pp. 179–197). New York: Wiley. Kessler, R. C. (1997). The prevalence of psychiatric comorbidity. In S. Wetzler & W. C. Sanderson (Eds.), Treatment strategies for patients with psychiatric comorbidity (pp. 23–48). New York: Wiley. Kessler, R. C. (2000). The epidemiology of pure and comorbid generalized anxiety disorder: A review and evaluation of recent research. Acta Psychiatrica Scandinavica, 406, 7–13. Kessler, R. C., Dupont, R. L., Berglund, P., & Wittchen, H.-U. (1999). Impairments in pure and comorbid generalized anxiety disorder and major depression at 12 months in two national surveys. American Journal of Psychiatry, 156, 1915–1923. Kessler, R. C., McGonagle, K. A., Zhao, S., Nelson, C. B., Hughes, M., Eshleman, S., Wittchen, H.-U., & Kendler, K. S. (1994). Lifetime and 12–month prevalence of DSM-III-R psychiatric disorders in the United States: Results from the National Comorbidity Survey. Archives of General Psychiatry, 51, 8–19. Kessler, R. C., Nelson, C. B., McGonagle, K. A., Liu, J., Swartz, M. S., & Blazer, D. G. (1996). Comorbidity of DSM-III-R major depressive disorder in the general population: Results from the U.S. National Comorbidity Survey. British Journal of Psychiatry, 168, 17–30. Kessler, R. C., & Price, R. H. (1993). Primary prevention of secondary disorders: A proposal and agenda. American Journal of Community Psychology, 21, 607–633. Lin, E., Katon, W., von Korff, M., Bush, T., Lipscomb, P., Russo, J., & Wagner, E. (1991). Frustrating patients: Physician and patient perspectives among distressed high users of medical services. Journal of General Internal Medicine, 6, 241–246. Maier, W., Gansicke, M., Freyberger, H. J., Linz, M., Heun, R., & Lecrubier, Y. (2000). Generalized anxiety disorder (ICD-10) in primary care from a cross-cultural perspective: A valid diagnostic entity? Acta Psychiatrica Scandinavia, 101, 29–36. Mancuso, D. M., Townsend, M. H., & Mercante, D. E. (1993). Long-term follow-up of generalized anxiety disorder. Comprehensive Psychiatry, 34, 441–446. Noyes, R., Holt, C. S., & Woodman, C. L. (1996). Natural course of anxiety disorders. In M. Mavissakalian & R. F. Prien (Eds.), Long-term treatments of anxiety disorders (pp. 1–48). Washington, DC: American Psychiatric Press. Noyes, R., Jr., Woodman, C., Garvey, M. J., Cook, B. L., Suelzer, M., Clancy, J., & Anderson, D. J. (1992). Generalized anxiety disorder vs. panic disorder: Distinguishing characteristics and patterns of comorbidity. Journal of Nervous and Mental Disease, 180, 369–379. Offord, D. R., Boyle, M. H., Campbell, D., Goring, P., Lin, E., Wrong, M., & Racine, A. (1996). One year prevalence of psychiatric disorder in Ontarians 15 to 64 years of age. Canadian Journal of Psychiatry, 41, 559–563. Offord, D. R., Campbell, D., Cochrane, J., Goering, P. N., Lin, E., Rhodes, A., & Wong, M. (1994). Mental health in Ontario: Selected findings from the Mental Health Supplement to the Ontario Health Survey. Toronto: Queen’s Printer for Ontario. Olfson, M., Fireman, B., Weissman, M. M., Leon, A. C., Sheehan, D. V., Kathol, R. G.,

Epidemiology

49

Hoven, C., & Farber, L. (1997). Mental disorders and disability among patients in a primary care group practice. American Journal of Psychiatry, 154, 1734–1740. Olfson, M., Kessler, R. C., Berglund, P. A., & Lin, E. (1998). Psychiatric disorder onset and first treatment contact in the United States and Ontario. American Journal of Psychiatry, 155, 1415–1422. Ormel, J., Van den Brink, W., Koeter, M. W., Giel, R., Van Der Meer, K., Van De Willige, G., & Wilmink, F. W. (1990). Recognition, management and outcome of psychological disorders in primary care: A naturalistic follow-up study. Psychiatric Medicine, 20, 909–923. Ormel, J., Von Korff, M., Ustun, B., Pini, S., Korten, A., & Oldehinkel, T. (1994). Common mental disorders and disability across cultures: Results from the WHO Collaborative Study on Psychological Problems in General Health Care. Journal of the American Medical Association, 272, 1741–1748. Reich, J. (1993). Distinguished mixed anxiety/depression from anxiety and depressive groups using the family history method. Comprehensive Psychiatry, 34, 285–290. Reich, J. (1995). Family psychiatric histories in male patients with generalized anxiety disorder and major depressive disorder. Annals of Clinical Psychiatry, 7, 71–78. Robins, E., & Guze, S. B. (1970). Establishment of diagnostic validity in psychiatric illness: Its application to schizophrenia. American Journal of Psychiatry, 126, 983– 987. Rogers, M. P., Warshaw, M. G., Goisman, R. M., Goldenberg, I., Rodriguez-Villa, F., Mallya, G., Freeman, S. A., & Keller, M. B. (1999). Comparing primary and secondary generalized anxiety disorder in a long-term naturalistic study of anxiety disorders. Depression and Anxiety, 10, 1–7. Roy-Byrne, P. P. (1996). Generalized anxiety and mixed anxiety–depression: Association with disability and health care utilization. Journal of Clinical Psychiatry, 57, 86–91. Salzman, C., Goldenberg, I., Bruce, S. E., & Keller, M. B. (2001). Pharmacologic treatment of anxiety disorders in 1989 versus 1996: Results from the Harvard/Brown Anxiety Disorders Research Program. Journal of Clinical Psychiatry, 62, 149–152. Schonfeld, W. H., Verboncoeur, C. J., Fifer, S. K., Lipschutz, R. C., Lubeck, D. P., & Bueschinf, D. P. (1997). The functioning and well-being of patients with unrecognized anxiety disorders and major depressive disorder. Journal of Affective Disorders, 43, 105–119. Skodal, A. E., Schwartz, S., Dohrenwend, B. P., Levav, I., & Shrout, P. E. (1994). Minor depression in a cohort of young adults in Israel. Archives of General Psychiatry, 51, 542–551. Spitzer, R. L., Kroenke, K., Linzer M., Hahn, S. R., Williams, J. B., deGruy, F. V., III, Brody, D., & Davies, M. (1995). Health-related quality of life in primary care patients with mental disorders. Journal of the American Medical Association, 274, 1511–1517. Stewart, A. L., Hays, R. D., & Ware, J. E., Jr. (1988). The MOS short-form general health survey: Reliability and validity in a patient population. Medical Care, 26, 724–735. Ustun, T. B., & Sartorius, N. (Eds.). (1995). Mental illness in general health care: An international study. Chichester, UK: Wiley. Wang, P. S., Berglund, P., & Kessler, R. C. (2000). Recent care of common mental disorders in the United States: Prevalence and conformance with evidence-based recommendations. Journal of General Internal Medicine, 15, 284–292. Warshaw, M. G., Keller, M. B., & Stout, R. L. (1994). Reliability and validity of the longi-

50

GAD IN THE CLINIC AND IN THE COMMUNITY

tudinal interval follow-up evaluation for assessing outcome of anxiety disorders. Journal of Psychiatric Research, 28(6), 531–545. Wittchen, H.-U., Zhao, S., Kessler, R. C., & Eaton, W. W. (1994). DSM-III-R generalized anxiety disorder in the National Comorbidity Survey. Archives of General Psychiatry, 51, 355–364. World Health Organization (WHO). (1990). International classification of diseases and related health problems (10th rev.): Classification of mental and behavioral disorders. Diagnostic criteria for research. Geneva: Author. Yonkers, K. A., Dyck, I. R., Warshaw, M., & Keller, M. B. (2000). Factors predicting the clinical course of generalized anxiety disorder. British Journal of Psychiatry, 176, 544–549. Yonkers, K. A., Massion, A., Warshaw, M., & Keller, M. B. (1996). Phenomenology and course of generalized anxiety disorder. British Journal of Psychiatry, 168, 308–313. Zuckerman, M. (1991). Psychobiology of personality. New York: Cambridge University Press.

GAD IN From A nxious THE CLINIC Temperament AND IN to THE Disorder COMMUNITY

CHAPTER 3

From Anxious Temperament to Disorder An Etiological Model JENNIFER L. HUDSON RONALD M. RAPEE

Despite the recent increase in the theoretical and empirical literature devoted to generalized anxiety disorder (GAD), surprisingly few publications have examined the potential etiological pathways to the disorder. Instead, the focus has been on the nature and treatment of GAD (Dugas, 2000). For a fuller understanding of the disorder, a complete picture that includes etiology is required. An awareness of the pathways to the disorder will, in turn, have important implications for the prevention and treatment of GAD. The current chapter reviews the available evidence regarding the development of GAD and presents a model that can be used as a guide for future etiological research.1 Of course, this model must be considered preliminary, as etiological research in the anxiety disorders remains in its infancy. Further limiting this quest is the small number of studies that have specifically examined the development of GAD. Most relevant research has either focused on other anxiety disorders (see Hudson & Rapee, 2000) or taken a broader perspective, examining the development of the anxiety disorders as a whole. 1

The model presented here is a modification of previously presented ideas (see Rapee, 2001).

51

52

GAD IN THE CLINIC AND IN THE COMMUNITY

Much of the research on GAD has suffered from the frequent and major changes in nosology across the editions of the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders (DSM; see Mennin, Heimberg, & Turk, Chapter 1, this volume). As a result, the few studies that have examined specific etiological factors in GAD have used a variety of different criteria, making it difficult to obtain a cohesive understanding of the ever-changing construct that is GAD. Despite these nosological changes, one view of GAD is that it is synonymous with elevated levels of trait anxiety (Rapee, 1991). Other researchers have also argued this point, suggesting that GAD represents a more fundamental type of anxiety that may underlie the other anxiety disorders (Barlow, 2002; Brown, Barlow, & Liebowitz, 1994). As a result, the model that we present is considered applicable to anxiety disorders in general. As etiological research advances, a more refined model that incorporates specific pathways to the individual disorders may be possible. The pathways to the anxiety disorders and more specifically to GAD are likely to be varied. Thus the model presented allows for a complex array of potential pathways. The research examining these potential pathways is explored. This chapter covers (1) genetic factors; (2) temperamental factors; and (3) environmental inf luences (including environmental support of avoidance, the effects of the social environment, and external environmental events).

GENETIC CONTRIBUTION Perhaps the most widely studied etiological component of anxiety is genetics. There has been a multitude of studies consistently showing that there is a significant genetic contribution to the variance in anxiety symptomatology and disorders (Andrews, Stewart, Allen, & Henderson, 1990; Andrews, Stewart, Morris-Yates, Holt, & Henderson, 1990; Jardine, Martin, & Henderson, 1984; Kendler, Neale, Kessler, Heath, & Eaves, 1992; Torgersen, 1983; Tyrer, Alexander, Remington, & Riley, 1987). Overall, these studies have suggested that genetics account for approximately 30–40% of the variance in anxiety symptomatology and disorders. Evidence from twin studies has tended to suggest that what is inherited is a general predisposition toward anxiety and depression rather than any specific genetic factor (Andrews, Stewart, Morris-Yates, et al., 1990; Torgersen, 1983). For example, Andrews, Stewart, Morris-Yates, and colleagues (1990) carried out a large twin study of anxiety and depressive disorders. There was no evidence for the specific heritability of individual anxiety disorders; rather, the data supported a notion that what is inherited is a general propensity toward neurosis. A more recent child study examining covariance between anxious and depressive symptoms in a sam-

From Anxious Temperament to Disorder

53

ple of 490 twin pairs aged 8–16 years also showed that shared genetics accounted for a large proportion of the variance (Eley, 1997). A slightly different picture has emerged from two twin studies of GAD, major depression, and other anxiety disorders, suggesting the possibility of two genetic factors (Kendler et al., 1995; Scherrer et al., 2000). Kendler and colleagues (1995) assessed 2,163 female twins for the presence of psychiatric disorders and showed that GAD and major depression were largely inf luenced by the same genetic factor (Factor 1). The study also showed a shared genetic risk for panic disorder, phobia, and bulimia nervosa (Factor 2). Although anxiety and depressive disorders loaded (to some extent) on both factors, the results indicated that there may be two somewhat separate genetic pathways for these disorders. That is, an individual may inherit a general predisposition toward (1) GAD and depression or (2) panic and phobia. In a similar study of male–male twin pairs, Scherrer and colleagues (2000) examined the degree to which liability for GAD and panic disorder was due to common genetic factors. This analysis showed that whereas 22.6% of the variance in symptoms was shared between GAD and panic disorder, a unique genetic inf luence (i.e., not shared with GAD) accounted for 21.2% of the variance in liability for panic disorder. Although the authors argued that it made no difference to the final model, diagnosis of GAD was based on liberal criteria, which included worry for at least 1 month, rather than the more conservative DSM-IV (American Psychiatric Association, 1994) criterion of 6 months. The 1-month criterion was used because of the low prevalence rates of GAD in the sample when the 6-month criterion was employed. Although the findings of Kendler and colleagues (1995) and Scherrer and colleagues (2000) suggest the presence of multiple genetic factors, their findings also provide continued support for a general genetic vulnerability toward anxiety and depression. At this stage, it would be premature to draw definite conclusions about the existence of more than one distinct genetic factor. Nevertheless, the evidence points to a general predisposition toward anxiety and depression rather than a specific heritability for GAD. Evidence from family studies of anxiety suggests that the environment, not genetics, is primarily responsible for the development of a specific anxiety disorder. In addition to providing evidence for a strong familial component to anxiety, the results of family studies suggest that this component may be largely disorder-specific. For example, Noyes, Clarkson, Crowe, Yates, and McChesney (1987) showed that first-degree relatives of individuals with GAD had an increased risk for GAD, but not panic disorder, in comparison to control families. Family studies of other anxiety disorders have also produced evidence for familial transmission of specific disorders (Fyer, Mannuzza, Chapman, Martin, & Klein, 1995; Last, Hersen, Kazdin, Orvaschel, & Perrin, 1991). Although genetics may

54

GAD IN THE CLINIC AND IN THE COMMUNITY

be heavily involved in an increased general risk for anxiety and depression, the development of specific disorders may be largely the result of the environment (Eley, 1997).

THE ROLE OF TEMPERAMENT One area of research that may help to answer the question of what is inherited is the study of temperament. “Temperament,” defined as an “intrinsic behavioral characteristic of a child” (Sanson, Prior, Garino, Oberklaid, & Sewell, 1987, p. 97), is likely to represent inherited qualities to some extent. Although definitions of temperament vary among theorists, it is generally agreed that temperament has a biological and genetic basis (Derryberry & Rothbart, 1984; Prior, Sanson, Oberklaid, & Northam, 1987; see Prior, 1992). Examining the link between certain temperamental patterns and later anxiety disorders may have important implications for understanding the manifestation of genetic heritability. Several researchers have identified dimensions of temperament. For example, Thomas and Chess (1985) identified three different patterns of temperament: (1) “easy” temperament, (2) “difficult” temperament, and (3) “slow-to-warm-up” temperament. For example, a classification of a difficult temperament was given to children who displayed a negative approach response to new situations, irregular eating and sleeping, frequent crying and tantrums, and slowness in adapting to change. Another example of dimensions of temperament comes from research using the Revised Infant Temperament Questionnaire (Carey & McDevitt, 1978). This questionnaire assesses nine temperamental factors: “approach,” “activity/ reactivity,” “food fussiness,” “rhythmicity,” “cooperation/manageability,” “placidity,” “threshold,” “irritability,” and “persistence.” Buss and Plomin (1984) also refer to three factors of temperament: “sociability,” “activity,” and “emotionality.” One area of temperament research that has gained much attention in recent times is the study of “behavioral inhibition” (BI). BI is a temperamental category defined in terms of reactions of withdrawal, wariness, avoidance, and shyness in novel situations, and is assessed following a series of laboratory observations of the child’s behavior in novel situations (Garcia Coll, Kagan, & Reznick, 1984; Reznick et al., 1986). Kagan (1989) suggests that approximately 15% of a normal sample of infants may be classified as exhibiting BI. Of relevance to the current review is the degree to which these temperamental dimensions and categories are related to later anxiety symptoms and disorders. Research that has examined this potential link has been carried out by examining (1) retrospective reports of anxious adults regarding their temperament as children; (2) temperament and anxiety

From Anxious Temperament to Disorder

55

symptoms in infants and children; (3) temperament in children at risk of developing anxiety disorders; or (4) temperament and its longitudinal association with later anxiety symptoms and disorders. Again, much of this research, with the exception of some of the BI studies, has broadly examined anxiety symptoms and disorders rather than GAD specifically.

Retrospective Reports from Anxious Adults The retrospective report of anxious adults has indicated that anxious behaviors may be identifiable from an early age. For example, adults with GAD commonly report being anxious all of their lives (Rapee, 1991). Studies of other anxiety disorders have also shown that anxious adults report higher levels of anxiety symptoms or anxiety-related behaviors as children in comparison to control groups (Lipsitz et al., 1994; Rapee & Melville, 1997).

Temperament and Anxiety in Children In research at the Child and Adolescent Anxiety Clinic at Macquarie University, mothers of clinically anxious children (n = 95) were more likely to report retrospectively that their children were “difficult” in the first year of life (crying, difficulties sleeping, pain, and gas) than mothers of children who were not clinically anxious (n = 75) (Rapee & Szollos, 1997). Differences between these groups were also evident in later years, with the mothers of anxious children reporting that their children had more fears in the second year of life, greater difficulties settling into day care and school, and greater difficulties adjusting to babysitters. Interestingly, children with specific anxiety disorders did not differ on these measures.

Temperament Studies of At-Risk Children Several studies have examined temperament in children at risk for developing an anxiety disorder. Merikangas, Avenevoli, Dierker, and Grillon (1999) measured temperament and psychophysiology in children of parents with an anxiety disorder. Children of anxious parents were not shown to differ from children of nonanxious parents on questionnaire measures of temperament. However they did differ on several psychophysiological measures that may ref lect differences in arousal. Thus it is possible that the differences in psychophysiology may ref lect a difference in vulnerability to anxiety in these children. Slightly different results were found in an examination of children of parents with panic disorder and agoraphobia and/or major depressive disorder (Rosenbaum et al., 1988). Children of parents with panic disorder and agoraphobia with or without comorbid depression were more

56

GAD IN THE CLINIC AND IN THE COMMUNITY

likely to be classified as BI than children of parents without panic disorder and agoraphobia.

Longitudinal Studies One of the most informative methods of examining the link between temperament and anxiety disorder is longitudinal research. Several studies have examined temperament and anxiety using this method (Biederman et al., 1993, 2001; Caspi, Henry, McGee, Moffitt, & Silva, 1995; Rende, 1993; Rubin, 1993). For example, Rende (1993) followed 1- to 4-year-old children until the age of 7 years. Mothers’ reports of temperament were assessed with the Emotionality, Activity, and Sociability Scale (Buss & Plomin, 1984), and mothers’ reports of the child’s internalizing and externalizing symptoms were assessed with the Child Behavior Checklist (CBCL; Achenbach & Edelbrock, 1979). Supporting the notion of a temperamental component in the development of anxiety, the study showed that anxiety and depressive symptoms at age 7 were related to (1) higher levels of emotionality in infancy and early childhood for both boys and girls, and (2) lower levels of sociability for girls only. Perhaps the most impressive body of longitudinal data relates to BI. The findings from a 3-year follow-up of infants with and without BI suggest that an inhibited child is significantly more likely than an uninhibited child to have DSM-III-R avoidant disorder of childhood, separation anxiety disorder, or agoraphobia at baseline measurement, and also more likely to develop avoidant disorder and separation anxiety disorder over the following 3-year period (Biederman et al., 1993). Children with stable BI over the 3-year period were more likely than children with unstable BI or without BI to develop an anxiety disorder. This result suggests that the stability of BI may be an important factor in the development of anxiety. In another study, 164 children identified at 4 months of age as either “high-reactive” (high levels of motor activity, high levels of crying) or “lowreactive” (low levels of motor activity/crying) were followed for 7 years (Kagan, Snidman, Zentner, & Peterson, 1999). In support of the previous findings, a greater number of high-reactive infants than of low-reactive infants (45% vs. 15%) exhibited anxious symptoms at age 7. Taken together, the results of retrospective and prospective studies support a link between early temperamental styles and anxiety. More specifically, temperamental styles of high arousal, emotionality, and BI are likely to place a child at risk for developing later anxiety symptoms and disorders. Although research has identified temperamental factors associated with later anxiety, it is important to note that not all temperamentally vulnerable children develop an anxiety disorder. This finding suggests that there must be other factors—probably environmental—that affect whether a vulnerable child ultimately develops an anxiety disorder. There

From Anxious Temperament to Disorder

57

has been much focus in the literature on potential environmental factors involved in the development of anxiety disorders.

ENVIRONMENTAL INFLUENCES In addition to providing information on heritability, behavior-genetic studies provide evidence for the instrumental role of the environment in the development of anxiety disorders. Results from adult twin studies indicate a fundamental role for nonshared environmental factors (experiences specific to the individual) in the development of anxiety disorders, while shared environmental factors (experiences common to siblings in a family) have been shown to play a minimal role (Jardine et al., 1984; Kendler et al., 1995). Some researchers have argued that the family is only likely to be important in the development of psychopathology when it is part of a child’s nonshared environment (Pike & Plomin, 1996; Rowe, 1997). In contrast, the results emerging from several recent twin studies of children demonstrate that shared environmental factors account for a significant amount of variance in both anxiety symptoms and disorders (Edelbrock, Rende, Plomin, & Thompson, 1995; Eley, 1997; Thapar & McGuffin, 1995; Topolski et al., 1997). In fact, in one study, shared environment was the only significant predictor of variance in self-reported anxiety symptoms in a sample of adolescent twins recruited from the community (Thapar & McGuffin, 1995). Although these results may seem to portray an exaggerated account of the role of shared environment, the overall results from child twin studies indicate a more important role for shared environment than is indicated by adult studies. Recently, however, Kendler (2001) has argued that two limitations of the twin study methodology may cause the results to underrepresent the role of the family environment. First, Kendler has argued that the twin modeling studies define the “family environment” as those experiences shared between siblings that contribute to similarity. When a child responds to a family event (such as divorce or parental discipline) differently from his or her sibling, this would be defined by twin studies as part of the child’s individual environment and not as part of the shared environment. Second, Kendler argues that twin studies have limited statistical power to assess the family environment, and large samples are required to detect even a moderate effect. Thus the twin modeling methodology may not provide an accurate assessment of the family environment. So in what way may one’s environment shape the development of GAD? Three sources of environmental inf luence that may be of importance in the development of anxiety disorders are (1) environmental support of avoidance, (2) transmission of threat and coping information, and (3) external environmental events.

58

GAD IN THE CLINIC AND IN THE COMMUNITY

Environmental Support of Avoidance Avoidance of threatening stimuli has been clearly identified as an integral factor in the maintenance of anxiety disorders. It is our belief that environmental support of avoidance behavior promotes the development of anxiety disorders by further fueling an anxious vulnerability. An inhibited child may select from his or her environment components that foster the child’s tendency to avoid novel/threatening stimuli. In turn, this continual avoidance may further shape the child’s temperamental vulnerability to pathology. In the early years, a child’s parents form the primary environment and therefore the greatest opportunity for supporting this avoidant style of interacting with the world. Through the years, the inf luence of siblings and peers may become increasingly important.

Parental Support of Avoidance Despite some behavior geneticists’ argument that the family environment is not important in the development of anxiety disorders, the potential role of parenting in the development of psychopathology has received much attention in the literature. In a review of parenting and the anxiety disorders, Rapee (1997) concluded that despite the multitude of methodological problems limiting the research in this area, there are surprisingly consistent findings linking controlling and overprotective parenting with the anxiety disorders. The ultimate consequence of these parenting behaviors may be the limitation of a child’s interaction with the environment. That is, a parent’s behavior may support a child’s avoidance. Although much of this research has been retrospective in nature, several studies of anxious children and their parents have allowed a more thorough testing of this association (e.g., Barrett, Rapee, Dadds, & Ryan, 1996; Chorpita, Albano, & Barlow, 1996; Hirshfeld, Biederman, Brody, Faraone, & Rosenbaum, 1997; Hudson & Rapee, 2001; Siqueland, Kendall, & Steinberg, 1996). Barrett and colleagues (1996) examined the responses of anxious children aged 7–14 years to a hypothetical situation of ambiguous threat before and after a discussion of the situation with their parents. Anxious children reported more avoidant coping responses after the discussion of the situation with their parents than before. In contrast, children with oppositional defiant disorder and control children showed a decrease in avoidant responses following the family discussion. Further analysis of the family discussion revealed that parents of anxious children supported avoidant responses by their child to the situation (Dadds, Barrett, Rapee, & Ryan, 1996). To further examine the parenting of anxious children, we designed a task that would enable the measurement of involvement and help-giving

From Anxious Temperament to Disorder

59

behaviors in mothers of anxious children (Hudson & Rapee, 2001). Children were given a moderately stressful puzzle task, and their mothers’ behavior during the task was assessed. Clear differences were identified in degree of involvement between mothers of anxious children and mothers of nonanxious children. Mothers of children with anxiety disorders gave more help and were more intrusive with their help during the task than were mothers of nonanxious children. These results further indicate that parents of anxious children may be attempting to prevent the children from experiencing distress by rushing in to help, rather than allowing the children to solve the problem on their own. Using a similar design, we examined parent–child interactions in the siblings of anxious children in order to determine whether overinvolved parenting was specific to the anxious children. We predicted that nonanxious siblings of anxious children would not experience the same amount of parental involvement during the task as the anxious children would. Our hypotheses were developed from theories of “child effects” on parenting; that is, parenting behaviors do not occur in isolation, but rather are inf luenced to some extent by a child’s temperament (e.g., Bell, 1977). Although the preliminary results of this study (Hudson & Rapee, 1998) supported our hypothesis, the final results showed a nonsignificant difference in parental involvement during the task for anxious children and their siblings (Hudson & Rapee, 2002). Compared to mothers of nonanxious children, mothers of anxious children were more involved during interactions with both the anxious children and the siblings of the anxious children. If, as the results of this study suggest, parental overinvolvement does not occur specifically in response to an anxious child, it is possible that factors such as parental anxiety or personality contribute to the development of an overprotective, overinvolved parenting style (Levy, 1943; Parker, 1983; Parker & Lipscombe, 1981). The majority of studies that have examined the relationship between parenting behavior and anxiety disorders have been limited by their inability to test causality and the direction of potential effect. One study that has employed longitudinal methods has indicated an interaction effect between parenting and temperament. In home observations of high- and low-reactive infants, Arcus (cited in Kagan, Snidman, Arcus, & Reznick, 1994) showed that the degree to which mothers held their infants when the infants did not need help predicted the children’s fear approximately a year later. However, this effect was found only for high-reactive infants. One possible explanation of these results is that parental overprotection on its own may not be sufficient to cause increased anxiety; however, when it occurs in the presence of inhibited temperament, it does so. Conclusions regarding the reciprocal relationship between temperament and parental behavior need to be postponed until further longitudinal inquiry of the relationship is carried out. Nevertheless, research so far

60

GAD IN THE CLINIC AND IN THE COMMUNITY

has established that parental overinvolvement—that is, behavior supporting a child’s restricted exposure to his or her environment—is linked with moderate consistency to the anxiety disorders in general. The research has not indicated a unique relationship between parenting and GAD. There has also been a substantial amount of literature and research on an infant’s attachment to his or her caregiver. According to Bowlby (1969/1974, 1973), an infant’s attachment or bond with his or her caregiver serves an evolutionary function to bring the infant into closer proximity with the caregiver. Ainsworth, Blehar, Waters, and Wall (1978) proposed that a securely attached infant uses the mother (or other primary caregiver) as a secure base from which to explore the environment. Ainsworth’s “Strange Situation” allows an infant’s behavior in relation to the mother to be categorized into one of four main attachment classifications: (1) avoidant, (2) secure, (3) ambivalent, or (4) disorganized (Ainsworth et al., 1978; Main & Solomon, 1986). An infant with an avoidant attachment tends to minimize distress during separation from and reunion with the caregiver. The baby resists contact with the caregiver throughout the situation and displays obvious avoidance of the caregiver on reunion. Parents of avoidant infants have been described as rejecting and insensitive to their children’s needs (Ainsworth et al., 1978). Secure infants seek proximity and contact with their caregivers, particularly on reunion. These infants may become distressed when mothers separate from them, but are easily soothed when the caregivers return. Mothers of secure infants have been described as sensitive to their children’s needs (Ainsworth et al., 1978). Conversely, children with ambivalent attachments maximize their distress following separation and reunion. They have extreme difficulty separating from their caregivers and have difficulty being soothed on reunion. Such a baby may demonstrate anger toward the caregiver or may appear passive. Mothers of ambivalent infants tend to be less rejecting than mothers of avoidant babies; nevertheless, they display less affectionate behavior than mothers of secure infants (Ainsworth et al., 1978). Children with disorganized attachments tend to have no systematic strategy for responding to their caregivers. This fourth category was introduced by Main and Solomon (1986) as a means of classifying infants whose behavior did not obviously fall into the other three categories. Infants with disorganized attachments are more likely to come from maltreated samples and are at greatest risk for social, biological, and developmental problems (Carlson, Cichetti, Barnett, & Braunwald, 1989; Cummings & Cicchetti, 1990; Main & Solomon, 1990). Only two studies have examined the relationship of attachment in infancy to the presence of anxiety disorders in childhood or adolescence (Shaw, Keenan, Vondra, Delliquadri, & Giovannelli, 1997; Warren, Huston, Egeland, & Sroufe, 1997). Warren and colleagues (1997) examined the attachment classifications of 172 infants at 12 months of age and

From Anxious Temperament to Disorder

61

followed them into adolescence. Only the first three attachment categories were examined (avoidant, secure, and ambivalent). Of the infants classified as having ambivalent attachments at 12 months, 28% were diagnosed with an anxiety disorder in adolescence, compared to 12% of securely attached infants and 16% of infants with avoidant attachments. Ambivalent attachment was associated with increased risk of developing an anxiety disorder in adolescence even after a mother’s anxiety and a child’s temperament were accounted for. Similarly, Shaw and colleagues (1997) followed a sample of 12-monthold infants from low-income families for 4 years. In this study, all four categories of attachment were assessed. Disorganized attachment at 12 months significantly predicted the child’s internalizing symptoms as measured by the CBCL (Achenbach & Edelbrock, 1979) at 5 years of age. However, on closer examination, the withdrawal items of the CBCL significantly correlated with disorganized attachment, whereas no significant association was found between attachment and items measuring depression or anxiety. These results suggest that although disorganized attachment may predict symptoms of withdrawal later in life, it does not necessarily predict symptoms of anxiety or depression. Maternal anxiety has also been identified as a factor contributing to the development of insecure attachment. For example, Manassis, Bradley, Goldberg, Hood, and Swinson (1994) showed that in a sample of 18 anxiety-disordered mothers of preschool-age children, 80% of the children were identified as insecurely attached (note that specific types of insecure attachments were not examined in this study). Thus children with anxious mothers may be more likely than children of nonanxious mothers to develop insecure attachments. As anxious children are more likely to have anxious parents (Last, Hersen, Kazdin, Francis, & Grubb, 1987), these results have important implications for children with anxiety disorders. Finally, Rapee and colleagues have recently shown that attachment style and temperament may interact in the production of anxious symptoms. Preschool children aged about 4 years were assessed both in the laboratory to determine inhibition and in a standard Strange Situation assessment with their mother (Shamir-Essakow, Ungerer, & Rapee, 2002). Anxiety diagnoses were determined via structured interview with the mothers. Ambivalent attachment style was associated with anxiety diagnoses in the children, but this was mediated by a strong correlation between ambivalent attachment and BI. On the other hand, avoidant attachment interacted with BI to predict anxiety diagnoses. Specifically, children who were showed both BI and avoidant attachment were much more likely to have received anxiety diagnoses than children in the other quadrants. In summary, the evidence from the attachment literature provides information that mother–child bonds at an early age may be predictive of later anxiety disorders. So far, one study has shown that infants with inse-

62

GAD IN THE CLINIC AND IN THE COMMUNITY

cure (ambivalent) attachments appear to be at greater risk of developing an anxiety disorder. However, it is possible that this is a result of shared variance with an inhibited temperament. On the other hand, there is emerging evidence that an avoidant attachment style may interact with inhibition to produce anxious symptomatology, although further longitudinal studies are needed.

Nonparental Support of Avoidance In a similar manner to spending time with parents, spending time with peers or siblings who are also anxious, introverted, or hesitant to take risks may further exacerbate an anxious child’s temperament by limiting those activities that could potentially allow habituation of his or her fears. This is an understudied area, and there is very limited empirical evidence to support these hypotheses. More recently, however, there has been a trend in the literature toward investigating the inf luence of peers in the development of personality and psychopathology. For example, some recent research has pointed to the possible importance of the peer group in maintaining attitudes related to weight concerns and dieting (Paxton, Schutz, Wertheim, & Muir, 1999). Some authors have even argued for the importance of the peer group to the virtual exclusion of parental inf luence (Harris, 1995). Although it is unlikely that peer inf luence exceeds the impact of other environmental factors, this is nevertheless an important area of investigation.

Transmission of Threat and Coping Information The environment can also be inf luential in the development of anxiety through the transmission of threat and coping information. This transmission may occur via modeling (Bandura & Rosenthal, 1966) or through the direct provision of threat or coping information. An individual may learn to be fearful or anxious in certain situations by watching the behavior of others. Although this method of learning has been less clearly defined for GAD, several studies using both animal and human subjects have demonstrated the pivotal role of modeling in the acquisition of phobias (Cook & Mineka, 1989; Muris, Steerneman, Merckelbach, & Meesters, 1996; Ollendick & King, 1991). For example, Cook and Mineka (1989) found that rhesus monkeys developed a fear of toy snakes by observing a videotape of other monkeys demonstrating fear responses to toy snakes. Furthermore, in a community study of 1,092 children, Ollendick and King (1991) showed that 56% of the children attributed the onset of their fears to vicarious conditioning experiences, and 39% of these children attributed the acquisition of their fear to being told stories about the feared object by parents, teachers, friends, or acquaintances. In a recent study, Gerull and Rapee (2002) investigated whether

From Anxious Temperament to Disorder

63

mothers’ modeling of fearful responses affected their infants’ approach– avoidance behavior. Mothers from the community and their infants aged 15–20 months were shown two novel threatening stimuli (a rubber snake and a spider). The mothers were instructed to reveal the object to the children and respond with a happy or fearful facial expression and tone. After 1- and 10-minute delays, the infants were presented with the stimuli again. This time, the mothers’ expressions were neutral. Observation of their behavior on the second and third trials showed that children were more likely to avoid the object when it was initially paired with a fearful expression. These results suggest that infants’ approach–avoidance behavior toward a novel object can be affected for up to 10 minutes following a single pairing with mothers’ facial expressions. The potential effect of repeated and extended modeling of anxious behavior in novel situations is likely to be more long-lasting. Even more dramatic effects may occur for inhibited children who already exhibit avoidance behavior in novel situations. The parents’ own anxiety may also contribute to the degree to which modeling shapes their children’s responses to novel objects. The more anxious a parent, the more likely he or she would be to demonstrate fearful or anxious behavior. The combination of an anxious parent and a vulnerable child is likely to augment the effect of modeling on the child’s behavior. Although these data refer to learning of specific fears, there is little doubt that GAD consists of fears of many specific situations and events. It might be speculated that if a young child is exposed to an anxious parent who displays fearfulness in response to a wide variety of cues, the child will eventually learn to generalize the ref lected fearfulness and will eventually internalize a more general rule, such as “The world is a dangerous place.” Children model their behavior not only on their parents’ behavior, but also on that of others, such as peers and siblings. As the children grow older, their behavior is more likely to be inf luenced by the opinions and behaviors of peers than by those of their parents. Adults too may model and learn from the anxious behavior of spouses/partners, peers, colleagues, or society. These sources of inf luence, though potentially important in shaping an anxious individual, have received far less attention in the literature than has the inf luence of parents.

External Environmental Events Contributing significantly to an understanding of the etiology of anxiety disorders is the awareness of the impact of external environmental events. This includes both the impact of stressful life events and the learning of particular associations through conditioning experiences. Positive experiences or learning of safety cues may also act as protective factors. The experience of stressful events during childhood has been clearly linked to an increased risk for the development of anxiety disorders (Tiet

64

GAD IN THE CLINIC AND IN THE COMMUNITY

et al., 1998). Retrospective reports of adults and children provide further support for this link (Brown & Harris, 1993; Brown, Harris, & Eales, 1993; Faravelli, Webb, Amboentti, Fonnescu, & Sessarego, 1985; Rapee & Szollos, 1997). For example, parents of clinically anxious children reported that their children experienced a greater number of stressful events in their early years, in comparison to the reports of mothers of nonanxious children (Rapee & Szollos, 1997). Specific research examining the occurrence of traumatic events in the lives of both adults and children with GAD also supports this link. Adults with GAD (clinical and undergraduate samples) were more likely than nonanxious adults to report exposure to a potentially traumatic event (Roemer, Molina, Litz, & Borkovec, 1997). Torgersen (1986) compared individuals with GAD to patients with panic disorder, showing that patients with GAD were more likely to report the death of a parent before age 16. In a sample of children with anxiety disorders, Manassis and Hood (1998) found that psychosocial adversity predicted the degree of impairment in children with GAD, but did not predict impairment in children with phobic disorders. Taken together, these findings provide preliminary evidence that traumatic events may play a significant role in the development of anxiety disorders, particularly GAD. Early adverse events have also been shown to have a significant and lasting effect on an individual’s physiology. More specifically, there is evidence that early stressful events can lead to increased activity of corticotropin-releasing factor neurons and to changes in the functioning of the hypothalamic–pituitary–adrenal axis (see Dienstbier, 1989; Heim & Nemeroff, 1999; Sapolsky, 1989). These changes in physiology may place a child at greater risk of developing an anxiety disorder by increasing his or her sensitivity to stress. Research with anxious adults has indicated that the number of stressful events is not related to the onset of anxiety. Rather, the impact of the negative events may be mediated by the individual’s vulnerability to anxiety. Rapee, Litwin, and Barlow (1990) asked clinically anxious adults (including those with GAD) and nonanxious adults to report on the number and impact of stressful life events in the 6 months preceding onset of their disorder. Raters who were unaware of participants’ anxiety status examined each of the stressors, providing an objective rating of the event’s negativity. There were no differences between the two groups in the number of stressful events or in the objective ratings of negativity. Despite this, clinically anxious adults reported a greater negative impact of the events than nonanxious adults, leading the authors to conclude that the impact of the negative life events was inf luenced by the individuals’ prior vulnerability to anxiety. That is, when an individual vulnerable to anxiety is faced with a negative life event, the event is likely to be experienced as more distressing. One factor that is considered integral in understanding the effects of

From Anxious Temperament to Disorder

65

stressful events is controllability (Chorpita & Barlow, 1998). Chorpita and Barlow (1998) hypothesize that early experience of reduced control over one’s environment will increase a child’s vulnerability to anxiety disorder. They propose that an environment of reduced control leads to an increased probability of interpreting future situations as uncontrollable. This may involve dramatic experiences of uncontrollability, such as chronic abuse or neglect, as well as chronically impoverished environments. But more commonly, it is likely to ref lect more subtle effects—such as overcontrolling parenting, in which the child is limited in his or her degree of decision making (e.g., see Rapee, 1997). Low perceptions of control are said to increase the individual’s underlying inhibition (Gray, 1987; Kagan, Reznick, & Snidman, 1987) and may also result in a general sense of lack of safety (e.g., Woody & Rachman, 1994). Some of the strongest support for this model comes from research with infant rhesus monkeys. Monkeys reared in an environment in which they had complete control over the delivery of food, water, and treats displayed significantly less fear and more exploratory behavior in novel environments than monkeys reared in situations of limited control (Mineka, Gunnar, & Champoux, 1986). These results support the hypothesis that control over one’s environment may be causally associated with lower levels of fearfulness. In summary, adverse life events, particularly uncontrollable events, may place an individual at greater risk of developing an anxiety disorder. Stressful, uncontrollable events early in life may alter the individual’s ability to respond to future stressors by increasing physiological arousal and sensitivity to future stressors. Although some stressors may be severe and traumatic enough to lead directly to the onset of an anxiety disorder, the impact of external environmental events is more likely to be the result of an interaction between the individual’s temperament and the stressor.

A MODEL OF THE DEVELOPMENT OF GAD From this review of the etiological research, it is obvious that many pathways may lead an individual to develop anxiety disorders such as GAD. These pathways are not yet fully understood, and so far only a partial picture can be offered. The model, presented in Figure 3.1, is partly speculative because of the limited research available on the etiology of GAD. The model depicts six main factors that have been discussed throughout this chapter: (1) genetic factors, (2) anxious vulnerability (manifested by temperament), (3) parental anxiety, (4) environmental support of avoidance, (5) transmission of threat and coping information, and (6) external environmental effects. It is clear that there is a strong genetic component contributing to the anxiety disorders, although the specific manifestation of these genetic fac-

66

GAD IN THE CLINIC AND IN THE COMMUNITY GENETIC FACTORS

PARENTAL ANXIETY

ANXIOUS VULNERABILITY AROUSAL AND EMOTIONALITY PROCESSING BIAS

AVOIDANCE

EXTERNAL ENVIRONMENTAL EVENTS

ENVIRONMENTAL SUPPORT OF AVOIDANCE

TRANSMISSION OF THREAT AND COPING INFORMATION

ANXIETY DISORDER

FIGURE 3.1. An etiological model of anxiety. From Rapee (2001). Copyright 2001 by Oxford University Press. Adapted by permission.

tors is not yet clear. What is clear is that GAD shares much of its genetic variance with both depression and the other anxiety disorders. There has been some evidence that there may also be a unique genetic contribution to individual anxiety disorders, although this is still far from clear. A child who inherits a genetic predisposition toward anxiety is likely to exhibit an anxious vulnerability, manifested partly as increased sensitivity, greater emotionality, and greater physiological arousal to threat. These temperamental variables are likely to contribute to the maintenance of the child’s vulnerability to anxiety. Greater emotionality and greater physiological arousal may give rise to an increased tendency to interpret situations as threatening, as well as a more avoidant coping style. An inhibited child who avoids novel stimuli is prevented from habituating to potentially fearful stimuli. The child’s restriction of exposure to the world increases the child’s sensitization to novel stimuli. This behavior may also lead to the limited development of social skills and coping skills, further contributing to the impact the disorder will have on the individual’s life. Initially, the child may be rewarded for avoidance of novel situations by the reduction of distress and attention from others in the environment. This maladaptive pattern may further serve to promote the child’s avoidance. In addition to behavioral avoidance, Borkovec, Alcaine, and Behar (Chapter 4, this volume) argue that worry, a key feature of GAD (Barlow, 2002; Rapee, 1991), ref lects a strategy of cognitive avoidance. Borkovec and colleagues argue that worry serves to suppress processing of threat information and lower physiological arousal. Despite the fact that GAD is

From Anxious Temperament to Disorder

67

characterized by above-average arousal, worry serves to restrict physiological arousal (Brown, Chorpita, & Barlow, 1998). It is possible that the child’s anxious temperament allows selection of behaviors from others in his or her environment (parents, siblings, teachers, peers) that provide support for the child’s style of interacting with the world. The relationship between the child’s anxious vulnerability and environmental support of avoidance is represented in Figure 3.1 with bidirectional arrows. The parents of a child with an anxious temperament may be more likely to become overinvolved with their child in an effort to reduce and prevent the child’s distress. This maladaptive pattern of parental overinvolvement in turn may reinforce the child’s vulnerability to anxiety by increasing the child’s perception of threat, reducing the child’s perceived control over threat, and ultimately increasing the child’s avoidance of threat. That is, parents who protect their child from stressful experiences or who take control in stressful situations may teach their child that the world is a dangerous place from which the child needs protection and over which he or she has no control. In terms of adult models of GAD, the child may thus be directed into an ongoing search for safety (Woody & Rachman, 1994) or may be left with a low illusion of control (Barlow, 2002). Conversely, parents who encourage their child’s independence and encourage their child to face difficult situations may reduce the likelihood of their child’s developing an anxiety disorder. Other significant persons (such as peers, siblings, spouse/partner, teachers, and colleagues) may also help to support and maintain this maladaptive style. In addition to supporting the individual’s natural avoidant coping style, various aspects of the social environment may provide an individual with new behaviors and attitudes. Foremost among these would be verbal instructions or modeling of anxious behaviors, which are likely to further shape an individual’s vulnerability toward disorder. This relationship is also likely to be of a bidirectional nature. An individual may model the anxious behavior of other significant persons in his or her environment (parents, peers, siblings) and learn how to behave in the face of fearful or ambiguous stimuli by observing these persons. To some extent, these socialization experiences may occur independently of the child’s anxious temperament. However, an anxious child is likely to choose an environment that increases the likelihood of being exposed to anxious modeling of behavior. For example, temperamentally anxious children may associate with anxious peers. In addition to transmission of information via modeling and verbalizations, cultural and social inf luences may provide the individual with anxious behaviors. Given the familial basis of anxiety (Last et al., 1991; Silverman, Cerny, & Nelles, 1988), parents of anxious children are also likely to be anxious and to exhibit a cognitive bias toward threat. Therefore, an anxious parent may be more likely to overprotect his or her child, due to an

68

GAD IN THE CLINIC AND IN THE COMMUNITY

increased perception of danger and an increased sensitivity to the child’s distress. The model depicts the contribution parental anxiety makes not only through the genetic transmission of anxiety, but also through the increased likelihood of modeling of anxious behavior and support of avoidant behavior. An individual who is vulnerable to anxiety may also experience stressful events that still further increase the individual’s vulnerability. Research has suggested that one of the integral factors in determining the impact of the stressful event is the controllability of the event: The more uncontrollable the environment, the greater the anxiety (Chorpita & Barlow, 1998). Furthermore, research has indicated that events related to danger, and not loss, are more likely to trigger anxiety than depression (Finlay-Jones & Brown, 1981). We predict that the experience of uncontrollable threat events interacts with the individual’s vulnerability to produce an episode of disorder. Given a sufficiently vulnerable individual, little or no environmental trigger may be necessary to produce an anxiety disorder. Similarly, given an environmental event of sufficient magnitude, even low-vulnerability individuals may develop a disorder (e.g., posttraumatic stress disorder). But in most cases, we expect a combination of more or less vulnerability interacting with more or less severe stressors to be involved in the development of an anxiety disorder. In the case of GAD, we would predict that the typical picture would be of greater input from temperamental vulnerability and little or no input from objectively dramatic life events. The distinction between vulnerability and disorder is a fine one. In fact, some studies have been unable to differentiate between individuals high in trait anxiety and clinically anxious individuals (e.g., see Ladouceur, Blais, Freeston, & Dugas, 1998). The difference between vulnerability and disorder, however, is likely to represent a difference in severity and interference of anxiety symptoms. With increasing severity and interference, the individual moves along a pathway from vulnerability to disorder. This progression may occur in response to specific or non-specific threat experiences; as a result, the onset of disorder may occur gradually with accumulating stressors, or dramatically in the case of a specific threat experience. The increase in “usual” levels of anxiety is accompanied by increased avoidance behavior, decreased perceptions of control and increased perceptions of threat, thereby maintaining the increased severity and interference of anxiety symptoms (see Rapee, 1991; Woody & Rachman, 1994) and maintaining the presence of disorder. The onset of disorder may also be inf luenced by the individual’s developmental level. For example, an inhibited, emotional child may not develop GAD until he or she reaches school age. At school, the child is presented with increased stressors, such as separation from parents, homework, performance situations, and formal schedules. There is a no-

From Anxious Temperament to Disorder

69

ticeable developmental shift in the foci of children’s fear from early childhood to adolescence (see Schniering, Hudson, & Rapee, 2000). In young children, fears tend to be more concrete, while in adolescence, anxiety is more likely to focus on social-evaluative concerns and more abstract worries. The child’s cognitive and emotional development is likely to have an impact on the transition from vulnerability to disorder. In conclusion, the model presented here is tentative. The etiological research has been far from conclusive about direct pathways to GAD. The model is intended as a guide for future research and highlights those areas that have received limited attention. The model also provides preliminary direction for research into the prevention of anxiety disorders. From the information available on the etiology of anxiety disorders, only a general model of anxiety disorders can truly be presented. It is hoped that in the future a clearer understanding of pathways to specific disorders will be possible. REFERENCES Achenbach, T. M., & Edelbrock, C. S. (1979). The Child Behavior Profile: 2. Boys aged 12–16 and girls aged 6–11 and 12–16. Journal of Consulting and Clinical Psychology, 47, 223–233. Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of attachment: A study of the Strange Situation. Hillsdale, NJ: Erlbaum. American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author. Andrews, G., Stewart, G. W., Allen, R., & Henderson, A. S. (1990). The genetics of six neurotic disorders: A twin study. Journal of Affective Disorders, 19, 23–29. Andrews, G., Stewart, G. W., Morris-Yates, A., Holt, P., & Henderson, A. S. (1990). Evidence for a general neurotic syndrome. British Journal of Psychiatry, 157, 6–12. Bandura, A., & Rosenthal, T. (1966). Vicarious classical conditioning as a function of arousal level. Journal of Personality and Social Psychology, 3, 54–62. Barlow, D. H. (2002). Anxiety and its disorders: The nature and treatment of anxiety and panic (2nd ed.). New York: Guilford Press. Barrett, P. M., Rapee, R. M., Dadds, M. R., & Ryan, S. (1996). Family enhancement of cognitive style in anxious and aggressive children. Journal of Abnormal Child Psychology, 24, 187–203. Bell, R. Q. (1977). Socialization findings reexamined. In R. Q. Bell & L. V. Harper (Eds.), Child effects on adults (pp. 53–84). Hillsdale, NJ: Erlbaum. Biederman, J., Rosenbaum, J. F., Bolduc-Murphy, E. A., Faraone, S. V., Chaloff, J., Hirshfeld, D. R., & Kagan, J. (1993). A 3–year follow-up of children with and without behavioral inhibition. Journal of the American Academy of Child and Adolescent Psychiatry, 32, 814–821. Biederman, J., Hirshfeld-Becker, D. R., Rosenabum, J. F., Herot, C., Friedman, D., Snidman, N., Kagan, J., & Faraone, S. V. (2001). Further evidence of association between behavioral inhibition and social anxiety in children. American Journal of Psychiatry, 158, 1673–1679. Bowlby, J. (1973). Attachment and loss: Vol. 2. Separation. London: Hogarth Press.

70

GAD IN THE CLINIC AND IN THE COMMUNITY

Bowlby, J. (1974). Attachment and loss: Vol. 1. Attachment. London: Hogarth Press. (Original work published 1969) Brown, G. W., & Harris, T. O. (1993). Etiology of anxiety and depressive disorders in an inner-city population: 1. Early adversity. Psychological Medicine, 23, 143–154. Brown, G. W., Harris, T. O., & Eales, M. J. (1993). Etiology of anxiety and depressive disorder in an inner-city population: 2. Comorbidity and adversity. Psychological Medicine, 23, 155–165. Brown, T. A., Barlow, D. H., & Liebowitz, M. R. (1994). The empirical basis of generalized anxiety disorder. American Journal of Psychiatry, 151, 1272–1280. Brown, T. A., Chorpita, B. F., & Barlow, D. H. (1998). Structural relationships among dimensions of the DSM-IV anxiety and mood disorders and dimensions of negative affect, positive affect, and autonomic arousal. Journal of Abnormal Psychology, 107, 179–192. Buss, A. H., & Plomin, R. (1984). Temperament: Early developing personality traits. Hillsdale, NJ: Erlbaum. Carey, W. B., & McDevitt, S. C. (1978). Revision of the Infant Temperament Questionnaire. Pediatrics, 68, 735–739. Carlson, V., Cicchetti, D., Barnett, D., & Braunwald, K. (1989). Disorganized/disoriented attachment relationships in maltreated infants. Developmental Psychology, 25, 525–531. Caspi, A., Henry, B., McGee, R. O., Moffitt, T. E., & Silva, P. A. (1995). Temperamental origins of child and adolescent behavior problems: From age three to age fifteen. Child Development, 66, 55–68. Chorpita, B. F., Albano, A. M., & Barlow, D. H. (1996). Cognitive processing in children: Relation to anxiety and family inf luences. Journal of Clinical Child Psychology, 25, 170–176. Chorpita, B. F., & Barlow, D. H. (1998). The development of anxiety: The role of control in the early environment. Psychological Bulletin, 124, 3–21. Cook, M., & Mineka, S. (1989). Observational conditioning of fear to fear-relevant versus fear-irrelevant stimuli in rhesus monkeys. Journal of Abnormal Psychology, 98, 448–459. Cummings, E. M., & Cicchetti, D. (1990). Toward a transactional model of relations between attachment and depression. In M. T. Greenberg, D. Cicchetti, & E. M. Cummings (Eds.), Attachment in the preschool years: Theory, research and interventions (pp. 339–372). Chicago: University of Chicago Press. Dadds, M. R., Barrett, P. M., Rapee, R. M., & Ryan, S. (1996). Family process and child anxiety and aggression: An observational analysis. Journal of Abnormal Child Psychology, 24, 715–734. Derryberry, D., & Rothbart, M. K. (1984). Emotion, attention and temperament. In C. Izard, J. Kagan, & R. Zajonc (Eds.), Emotions, cognition, and behavior (pp. 132–166). Cambridge, UK: Cambridge University Press. Dienstbier, R. A. (1989). Arousal and physiological toughness: Implications for mental and physical health. Psychological Review, 96, 84–100. Dugas, M. J. (2000). Generalized anxiety disorder publications: So where do we stand? Journal of Anxiety Disorders, 14, 31–40. Edelbrock, C., Rende, R., Plomin, R., & Thompson, L. A. (1995). A twin study of competence and problem behavior in childhood and early adolescence. Journal of Child Psychology and Psychiatry, 36, 775–785. Eley, T. C. (1997). General genes: A new theme in developmental psychopathology. Current Directions in Psychological Science, 6, 90–95. Faravelli, C., Webb, T., Amboentti, A., Fonnescu, F., & Sessarego, A. (1985). Prevalence

From Anxious Temperament to Disorder

71

of traumatic early events in 31 agoraphobic patients with panic attacks. American Journal of Psychiatry, 142, 1493–1494. Finlay-Jones, R., & Brown, G. W. (1981). Types of stressful life events and the onset of anxiety and depressive disorders. Psychological Medicine, 11, 803–815. Fyer, A. J., Mannuzza, S., Chapman, T. F., Martin, L. Y., & Klein, D. F. (1995). Specificity in familial aggregation of phobic disorders. Archives of General Psychiatry, 52, 564–573. Garcia Coll, C., Kagan, J., & Reznick, J. S. (1984). Behavioral inhibition in young children. Child Development, 55, 1005–1019. Gerull, F. C., & Rapee, R. M. (2002). Mother knows best: Effects of maternal modelling on the acquisition of fear and avoidance behaviour in toddlers. Behaviour Research and Therapy, 40, 279–287. Gray, J. A. (1987). The psychology of fear and stress (2nd ed.). Cambridge, UK: Cambridge University Press. Harris, J. R. (1995). Where is the child’s environment?: A group socialization theory of development. Psychological Review, 102, 458–489. Heim, C., & Nemeroff, C. B. (1999). The impact of early adverse experiences on brain systems involved in the pathophysiology of anxiety and affective disorders. Biological Psychiatry, 46, 1509–1522. Hirshfeld, D. R., Biederman, J., Brody, L., Faraone, S. V., & Rosenbaum, J. F. (1997). Associations between expressed emotion and child behavioral inhibition and psychopathology: A pilot study. Journal of the American Academy of Child and Adolescent Psychiatry, 36, 205–213. Hudson, J. L., & Rapee, R. M. (1998, July). Parenting of anxious children and their siblings. Paper presented at the triennial meeting of the World Congress of Behavioral and Cognitive Therapies, Acapulco, Mexico. Hudson, J. L., & Rapee, R. M. (2000). The origins of social phobia. Behavior Modification, 24, 102–129. Hudson, J. L., & Rapee, R. M. (2001). Parent–child interactions and the anxiety disorders: An observational analysis. Behaviour Research and Therapy, 39, 1411–1427. Hudson, J. L., & Rapee, R. M. (2002). Parent–child interactions in clinically anxious children and their siblings. Journal of Clinical Child and Adolescent Psychology, 31(4), 548–555. Jardine, R., Martin, N. G., & Henderson, A. S. (1984). Genetic covariance between neuroticism and the symptoms of anxiety and depression. Genetics Epidemiology, 1, 89–107. Kagan, J. (1989). Temperamental contributions to social behavior. American Psychologist, 44, 668–674. Kagan, J., Reznick, J. S., & Snidman, N. (1987). The physiology and psychology of behavioral inhibition in children. Child Development, 58, 1459–1473. Kagan, J., Snidman, N., Arcus, D., & Reznick, J. S. (1994). Galen’s prophecy: Temperament in human nature. New York: Basic Books. Kagan, J., Snidman, N., Zentner, M., & Peterson, E. (1999). Infant temperament and anxious symptoms in school age children. Development and Psychopathology, 11, 209–224. Kendler, K. S. (2001). Twin studies of psychiatric illness: An update. Archives of General Psychiatry, 58, 1005–1014. Kendler, K. S., Neale, M. C., Kessler, R. C., Heath, A. C., & Eaves, L. J. (1992). Major depression and generalized anxiety disorder: Same genes, (partly) different environments? Archives of General Psychiatry, 49, 716–722. Kendler, K. S., Walters, E. E., Neale, M. C., Kessler, R. C., Heath, A. C., & Eaves, L. J.

72

GAD IN THE CLINIC AND IN THE COMMUNITY

(1995). The structure of the genetic and environmental risk factors for six major psychiatric disorders in women: Phobia, generalized anxiety disorder, panic disorder, bulimia, major depression, and alcoholism. Archives of General Psychiatry, 52, 374–383. Ladouceur, R., Blais, F., Freeston, M. H., & Dugas, M. J. (1998). Problem solving and problem orientation in generalized anxiety disorder. Journal of Anxiety Disorders, 12, 139–152. Last, C. G., Hersen, M., Kazdin, A., Francis, G., & Grubb, H. J. (1987). Psychiatric illness in the mothers of anxious children. American Journal of Psychiatry, 144, 1580– 1583. Last, C. G., Hersen, M., Kazdin, A., Orvaschel, H., & Perrin, S. (1991). Anxiety disorders in children and their families. Archives of General Psychiatry, 48, 928–935. Levy, D. M. (1943). Maternal overprotection. New York: Columbia University Press. Lipsitz, J. D., Martin, L. Y., Mannuzza, S., Chapman, T. F., Liebowitz, M. R., Klein, D. F., & Fyer, A. J. (1994). Childhood separation anxiety disorder in patients with adult anxiety disorders. American Journal of Psychiatry, 151, 927–929. Main, M., & Solomon, J. (1986). Discovery of an insecure–disorganized/disoriented attachment pattern. In T. B. Brazelton & M. W. Yogman (Eds.), Affective development in infancy (pp. 95–124). Norwood, NJ: Ablex. Main, M., & Solomon, J. (1990). Procedures for identifying infants as disorganized/disoriented during the Ainsworth Strange Situation. In M. Greenberg, D. Cicchetti, & E. M. Cummings (Eds.), Attachment in the preschool years: Theory, research and intervention (pp. 121–160). Chicago: University of Chicago Press. Manassis, K., Bradley, S., Goldberg, S., Hood, J., & Swinson, R. P. (1994). Attachment in mothers with anxiety disorders and their children. Journal of the American Academy of Child and Adolescent Psychiatry, 33, 1106–1113. Manassis, K., & Hood, J. (1998). Individual and familial predictors of impairment in childhood anxiety disorders. Journal of the American Academy of Child and Adolescent Psychiatry, 37, 428–434. Merikangas, K. R., Avenevoli, S., Dierker, L., & Grillon, C. (1999). Vulnerability factors among children at risk for anxiety disorders. Biological Psychiatry, 46, 1523–1535. Mineka, S., Gunnar, M., & Champoux, M. (1986). Control and early socioemotional development: Infant rhesus monkeys reared in controllable versus uncontrollable environments. Child Development, 57, 1241–1256. Muris, P., Steerneman, P., Merckelbach, H., & Meesters, C. (1996). The role of parental fearfulness and modeling of children’s fears. Behaviour Research and Therapy, 34, 265–268. Noyes, R., Jr., Clarkson, C., Crowe, R. R., Yates, W. R., & McChesney, C. M. (1987). A family study of generalized anxiety disorder. American Journal of Psychiatry, 144, 1019–1024. Ollendick, T. H., & King, N. J. (1991). Origins of childhood fears: An evaluation of Rachman’s theory of fear acquisition. Behaviour Research and Therapy, 29, 117– 123. Parker, G. (1983). Parental overprotection: A risk factor in psychosocial development. Sydney: Grune & Stratton. Parker, G., & Lipscombe, P. (1981). Inf luences on maternal protection. British Journal of Psychiatry, 138, 303–311. Paxton, S. J., Schutz, H. K., Wertheim, E. H., & Muir, S. L. (1999). Friendship clique and peer inf luences on body image concerns, dietary restraint, extreme weightloss behaviors, and binge eating in adolescent girls. Journal of Abnormal Psychology, 108, 255–266.

From Anxious Temperament to Disorder

73

Pike, A., & Plomin, R. (1996). Importance of nonshared environment factors for childhood and adolescent psychopathology. Journal of the American Academy of Child and Adolescent Psychiatry, 35, 560–570. Prior, M. (1992). Childhood temperament. Journal of Child Psychology and Psychiatry, 33, 249–279. Prior, M., Sanson, A., Oberklaid, F., & Northam, E. (1987). Measurement of temperament on one to three year old children. International Journal of Behavioural Development, 10, 121–132. Rapee, R. M. (1991). Generalized anxiety disorder: A review of clinical features and theoretical concepts. Clinical Psychology Review, 11, 419–440. Rapee, R. M. (1997). The potential role of childrearing practices in the development of anxiety and depression. Clinical Psychology Review, 17, 47–67. Rapee, R. M. (2001). The development of generalized anxiety. In M. W. Vasey & M. R. Dadds (Eds.), The developmental psychopathology of anxiety (pp. 481–503). New York: Oxford University Press. Rapee, R. M., Litwin, E. M., & Barlow, D. H. (1990). Impact of life events on subjects with panic disorder and on comparison subjects. American Journal of Psychiatry, 147, 640–644. Rapee, R. M., & Melville, L. F. (1997). Recall of family factors in social phobia and panic disorder: Comparison of mother and offspring reports. Depression and Anxiety, 5, 7–11. Rapee, R. M., & Szollos, A. (1997, November). Early life events in anxious children. Paper presented at the 31st annual meeting of the Association for Advancement of Behavior Therapy, Miami, FL. Rende, R. D. (1993). Longitudinal relations between temperament traits and behavioral syndromes in middle childhood. Journal of the American Academy of Child and Adolescent Psychiatry, 32, 287–290. Reznick, J. S., Kagan, J., Snidman, N., Gersten, M., Baak, K., & Rosenberg, A. (1986). Inhibited and uninhibited children: A follow-up study. Child Development, 54, 361– 368. Roemer, L., Molina, S., Litz, B. T., & Borkovec, T. D. (1997). Preliminary investigation of the role of previous exposure to potentially traumatizing events in generalized anxiety disorder. Depression and Anxiety, 4, 134–138. Rosenbaum, J. F., Biederman, J., Gersten., M., Hirshfeld, D. R., Meminger, S. R., Herman, J. B., Kagan, J., Reznick, J. S., & Snidman, N. (1988). Behavioral inhibition in children of parents with panic disorder and agoraphobia: A controlled study. Archives of General Psychiatry, 45, 463–470. Rowe, D. C. (1997). As the twig is bent?: The myth of child-rearing inf luences on personality development. Journal of Counseling and Development, 68, 606–611. Rubin, K. H. (1993). The Waterloo longitudinal project: Correlates and consequences of social withdrawal from childhood to adolescence. In K. H. Rubin & J. B. Asendorpf (Eds.), Social withdrawal, inhibition, and shyness in children (pp. 291– 314). Hillsdale, NJ: Erlbaum. Sanson, A., Prior, M., Garino, E., Oberklaid, F., & Sewell, J. (1987). The structure of infant temperament: Factor analysis of the Revised Infant Temperament Questionnaire. Infant Behavior and Development, 10, 97–104. Sapolsky, R. M. (1989). Hypercortisolism among socially subordinate wild baboons originates at the CNS level. Archives of General Psychiatry 46, 1047–1051. Scherrer, J. F., True, W. R., Xian, H., Lyons, M. J., Eisen, S. A., Goldberg, J., Lin, N., & Tsuang, M. T. (2000). Evidence for genetic inf luences common and specific to symptoms of generalized anxiety and panic. Journal of Affective Disorders, 57, 25–35.

74

GAD IN THE CLINIC AND IN THE COMMUNITY

Schniering, C. A., Hudson, J. L., & Rapee, R. M. (2000). Issues in the diagnosis and assessment of anxiety disorders in children and adolescents. Clinical Psychology Review, 20, 453–478. Shamir-Essakow, G., Ungerer, J., & Rapee, R. M. (2002). Attachment, behavioural inhibition, and anxiety problems in preschool children. Manuscript submitted for publication. Shaw, D. S., Keenan, K., Vondra, J. I., Delliquadri, E., & Giovannelli, B. S. (1997). Antecedents of preschool children’s internalizing problems: A longitudinal study of low-income families. Journal of the American Academy of Child and Adolescent Psychiatry, 36, 1760–1767. Silverman, W. K., Cerny, J. A., & Nelles, W. B. (1988). The familial inf luence in anxiety disorders: Studies on the offspring of patients with anxiety disorders. In B. B. Lahey & A. E. Kazdin (Eds.), Advances in clinical child psychology (pp. 223–248). New York: Plenum Press. Siqueland, L., Kendall, P. C., & Steinberg, L. (1996). Anxiety in children: Perceived family environments and observed family interaction. Journal of Clinical Child Psychology, 25, 225–237. Thapar, A., & McGuffin, P. (1995). Are anxiety symptoms in childhood heritable? Journal of Child Psychology and Psychiatry, 36, 439–447. Thomas, A., & Chess, S. (1985). The behavioral study of temperament. In J. Strelau & F. H. Farley (Eds.), The biological bases of personality and behavior: Vol. 1. Theories, measurement techniques and development (pp. 213–225). Washington, DC: Hemisphere. Tiet, Q. Q., Bird, H. R., Davies, M., Hoven, C., Cohen, P., Jensen, P. S., & Goodman, S. (1998). Adverse life events and resilience. Journal of the American Academy of Child and Adolescent Psychiatry, 37, 1191–1200. Topolski, T. D., Hewitt, J. K., Eaves, L. J., Silberg, J. L., Meyer, J. M., Rutter, M., Pickles, A., & Simonoff, E. (1997). Genetic and environmental inf luences on child reports of manifest anxiety and overanxious disorders: A community-based twin study. Behavior Genetics, 27, 15–28. Torgersen, S. (1983). Genetic factors in anxiety disorders. Archives of General Psychiatry, 40, 1085–1089. Torgersen, S. (1986). Childhood and family characteristics in panic and generalized anxiety disorders. American Journal of Psychiatry, 143, 630–632. Tyrer, P., Alexander, J., Remington, M., & Riley, P. (1987). Relationship between neurotic symptoms and neurotic diagnosis: A longitudinal study. Journals of Affective Disorders, 13, 13–21. Warren, S. L., Huston, L., Egeland, B., & Sroufe, L. A. (1997). Child and adolescent anxiety disorders and early attachment. Journal of the American Academy of Child and Adolescent Psychiatry, 36, 637–644. Woody, S., & Rachman, S. (1994). Generalized anxiety disorder (GAD) as an unsuccessful search for safety. Clinical Psychology Review, 14, 743–753.

PART II

Theoretical and Empirical Approaches

THEORETICAL Avoidance Theor AND y of Worry EMPIRICAL APPROACHES

CHAPTER 4

Avoidance Theory of Worry and Generalized Anxiety Disorder THOMAS D. BORKOVEC OSCAR M. ALCAINE EVELYN BEHAR

Worry is a pervasive human experience and is the defining characteristic of generalized anxiety disorder (GAD) (see Mennin, Heimberg, & Turk, Chapter 1, this volume). Determining what functions it serves in human anxiety may be critical to the development of therapeutic methods for ameliorating the suffering that it causes. Perhaps its most significant function resides in the likely use of the process as an internal avoidance response. At one level of analysis, the claim that worry is a cognitive avoidance response to perceived threat would seem to be self-evident. If a human being is anticipating a significant problem, he or she will be highly motivated to solve (remove) the problem. Cognitive activity, including the ability to recall the past and think about or imagine possible futures and behaviors to adapt to such futures, is a highly sophisticated system for problem solving. In fact, modern cognitive theory posits that thought evolved precisely to anticipate the future and to identify ways to maximize good outcomes and minimize bad ones (e.g., McGuire & McGuire, 1991). Detecting threat creates a problem to be solved. Indeed, threat represents the most significant problem that organisms face, given that it relates to survival. Threat signals the need to find an effective way to avoid or minimize bad events. If behavioral avoidance cannot be employed to remove the threat and the problem remains unsolved, a solution to the threat 77

78

THEORETICAL AND EMPIRICAL APPROACHES

must still be found for the sake of survival, and cognitive activity is the only device left available. Worry is thus an understandable cognitive attempt to solve the problem of possible future danger; it functions to avoid the occurrence of temporally distant bad events. To the extent that worry does indeed reduce aspects of the threat and/or is perceived to prevent bad events, worry should be negatively reinforced like any avoidance behavior. The extant literature suggests that worry does indeed function as a cognitive avoidance response (potentially in several different ways), and that the conditions necessary for negative reinforcement are frequently present. The purpose of the present chapter is to review empirical literature relevant to the potential avoidance functions of worry; to describe the significant consequences of that process; and to suggest areas of possible therapeutic innovation for reducing chronic worry and GAD, based on what is known about its nature and functions. It is first useful, however, to provide a brief overview of the historical context within which scientific research on worry emerged.

A HISTORICAL CONTEXT FOR THE EMERGENCE OF INTEREST IN WORRY The Two-Stage Theory of Anxiety From an evolutionary perspective, fear is advantageous; it facilitates fightor-f light responses to environments that threaten survival. Although fear can serve an adaptive function in this way, it can also be manifested in the pathological forms seen in the anxiety disorders, wherein the fear is excessive and/or leads to maladaptive behavior in response to situations that are not actually threatening. From a classic learning theory perspective, we humans develop anxiety as a response to nonthreatening situations because we have been punished in the past in the presence of those situations. Once such an emotional response is conditioned, we naturally seek to escape or avoid those situations because we do not wish to experience the punishment again. As long as we avoid the conditioned stimuli, however, extinction cannot occur, and the learned anxiety response will be maintained. This description is very familiar to the field of behavior therapy. A significant portion of its theoretical foundation, from which many of its exposure therapy methods have emerged, rests in Mowrer’s (1947) two-stage theory of fear: Classically conditioned acquisition of fear is followed by operantly conditioned avoidance of fear cues, resulting in fear maintenance due to a lack of unreinforced exposure to those conditioned stimuli. In the application of this perspective to human anxiety, motoric avoidance and external fear cues naturally received the greatest amount of attention in the early days of behavior therapy, given that field’s initial emphasis on observable behaviors and on the understanding and

Avoidance Theory of Worry

79

treatment of circumscribed phobic disorders for which clearly identifiable environmental triggers existed. Despite significant therapeutic advances in the treatment of several of the anxiety disorders, these two early emphases were not very useful for attempts to address diffuse anxiety, or what eventually became GAD in the third edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-III; American Psychiatric Association [APA], 1980). This disorder was not characterized by specific, anxiety-provoking triggers, and behavioral avoidance was not an obvious symptomatic feature. Without specific external fear cues to use in exposure, and without an approach response to encourage to counteract avoidance of such cues, traditional exposure techniques seemed to have little relevance.

The Role of Cognition and the Conditioned Response in Anxiety Process Two separate but related lines of thought were, however, emerging shortly after the beginnings of behavior therapy that ultimately provided a foundation for conceptual and therapeutic advances for GAD. First, with the advent of interest in cognitive processes potentially involved in psychological problems (Bandura, 1969), the possibility that humans might be capable of cognitively avoiding feared stimuli (and thus preserving anxious meanings despite repeated exposures and without behavioral avoidance) was eventually raised. In the most obvious example of this, the fact that phobic stimuli may be presented in daily life or in therapy does not guarantee that phobic individuals will attend to those stimuli; they may well avoid them by actively shifting their attention to something else. Second, with increasing awareness that the response-produced stimuli of the conditioned response might themselves be part of the conditioned stimulus complex, the possibility that exposure to internal anxiety responses might be important for effective and efficient exposure therapies was increasingly considered. Under this circumstance, avoidance of internal fear cues (either by covert or overt means) might well play a role in the maintenance of anxiety. The implications of the view that anxiety responses are part of the conditioned stimulus complex have historically been played out in various developments, such as the deliberate elicitation of fear in implosive therapy (Stampf l & Levis, 1967), the concept of “fear of fear” in agoraphobia (Goldstein & Chambless, 1978), and the use of interoceptive exposure in panic control treatment (Barlow, 1988). These emerging lines of thought regarding the role of cognitive processes in anxiety and the response-produced stimulus value of internal fear cues eventually culminated in Foa and Kozak’s (1986) theoretical notion that accessing the full fear structure stored in memory (including its stimulus, response, and meaning propositions, based on Lang’s [1985] bioinformational model of fear) and the provision of corrective information during repeated exposures are the central mechanisms of the extinction process.

80

THEORETICAL AND EMPIRICAL APPROACHES

Our own interest in cognitive avoidance dates back to the time of Thomas Borkovec’s dissertation, which compared systematic desensitization, implosive therapy, and an avoidance response placebo in the treatment of analogue snake phobia (Borkovec, 1972, 1974). Because the theoretical underpinnings of both these exposure therapies were grounded in classic learning theory, the ideal control condition for them would logically involve the provision of the same amount of exposure to feared hierarchy scenes, but with the addition of the imagination of an avoidance response to each scene presentation. Based on Mowrer’s (1947) theory, these are precisely the conditions assumed to occur in daily life exposures to anxiety-provoking stimuli that should maintain the anxiety. The fact that the avoidance response group failed to show outcome improvement on heart rate measures and displayed maintenance of heart rate arousal throughout the process of phobic scene exposures (in contrast to the declining heart rates of the two behavioral therapies) provided evidence that cognitive (imagined) avoidance, like behavioral avoidance, could function to preserve anxiety despite repeated confrontations with the feared situation. A subsequent investigation (Grayson & Borkovec, 1978) similarly found that imagining avoidance responses to phobic scenes resulted in greater progressive increases in subjective fear across hierarchy images than did a condition involving imagination of effective coping with the feared situation. Moreover, Grayson, Foa, and Steketee (1982) found that attention deployed toward or away from anxiety-provoking stimuli during exposure treatment significantly inf luences the rate of between-session extinction. These findings suggested a cognitive–perceptual extension of Mowrer’s two-stage theory (Borkovec, 1979), which posited that cognitive escape or avoidance might (1) contribute to the original development of anxious responding, (2) facilitate the maintenance of already established phobias, and (3) mitigate the otherwise ameliorating effects of repeated exposures occurring in the course of daily life or in therapy. The latter two effects were hypothetically due to the fact that such cognitive maneuvers would preclude the functional exposure necessary to promote extinction. Also contained within this theoretical position was an emphasis on the likelihood that the conditioned anxiety response itself might be a part of the conditioned stimulus complex, and therefore potentially a necessary ingredient for effective exposure treatment. By the late 1980s, DSM-III-R (APA, 1987) changed its diagnostic criteria for GAD, making a cognitive process (worry) its central defining characteristic (see Mennin et al., Chapter 1, this volume). Since that time, although some progress had been made in applying traditional cognitivebehavioral techniques (e.g., relaxation training, generic cognitive therapy, and self-control desensitization techniques) to the treatment of the disorder (see Leahy, Chapter 11, this volume), it also became apparent that the

Avoidance Theory of Worry

81

development of effective therapies for GAD might well depend upon a better, more specific understanding of the nature and functions of worry. And the extension of theory to include cognitive avoidance and the importance of accessing all elements of the fear structure provided one of the useful bases for the search for such an understanding.

Worry as a Cognitive Process Relevant to Anxiety Experimental interest in worry itself emerged mainly in the early 1970s within the test anxiety literature, where worry was defined as a cognitive component of anxiety distinct from its physiological elements. The fact that measures of worry predicted such things as test performance and grade point average, whereas measures of the physiological anxiety element did not, supported that distinction (see Deffenbacher, 1980). When Borkovec and colleagues began their experimental studies of worry, they were already convinced by this test anxiety research, by their interviews of individuals with chronic worry, and by their own introspective experiences that worry was closely related to fear and anxiety processes (see Borkovec, Robinson, Pruzinsky, & DePree, 1983). Indeed, empirical investigations found that anxiety was the predominant affective experience during worry and that worry, like fear, had incubating effects, generating an increase in negative thought intrusions after brief inductions compared to the decrease in intrusions after lengthy inductions (Borkovec, Robinson, et al., 1983; Pruzinsky & Borkovec, 1990). So Borkovec (1985) speculated that worry should also have some type of cognitive avoidance function, and his group was primed to recognize such functions, should they emerge in subsequent research. Any theory proposing the central role of an avoidance behavior in pathological anxiety necessarily assumes the importance of threat detection; it is the perception of threat that initiates the anxiety process. As MacLeod and Rutherford (Chapter 5, this volume) comprehensively describe, worry and GAD are indeed characterized by a pervasive bias toward threat detection and interpretation. Our chapter focuses on worry as one of the significant ways in which human beings respond to perceived threat.

COGNITIVE AVOIDANCE THROUGH THE WORRY PROCESS The Predominance of Thought Activity in the Worry Process The first empirical hint that worry may function as an avoidance response emerged from an investigation of clients with GAD and nonanxious controls, from whom thought samples were collected during periods of selfrelaxation and worry (Borkovec & Inz, 1990). During relaxation, the cli-

82

THEORETICAL AND EMPIRICAL APPROACHES

ents reported nearly equal amounts of thought and imagery, whereas the controls reported a predominance of imagery. During the subsequent worry period, thought predominated for both groups. Clients with GAD also rated the affective valence of their mentations as significantly more anxious than did the controls, even during the relaxation state. Importantly, both the thought–imagery ratios and the valence ratings normalized once the clients completed a trial of psychotherapy. This study and a replication (Freeston, Dugas, & Ladouceur, 1996) established that worry is phenomenologically experienced primarily as negative verbal/linguistic (as opposed to imaginal) activity. When we humans worry, we are talking to ourselves in anxious ways. Later psychophysiological evidence strengthened the conclusion of thought predominance; both GAD as a condition and worry as a state have been associated with increases in left frontal cortical activation (Borkovec, Ray, & Stöber, 1998; Carter, Johnson, & Borkovec, 1986; Heller, Nitschke, Etienne, & Miller, 1997; Tucker, Antes, Stenslie, & Barnhardt, 1978; Tyler & Tucker, 1982). Although the fact that imagery declined from relaxed states among nonanxious people to worrisome states among clients with GAD in the Borkovec and Inz (1990) study might merely have indicated that the excessive thought contained in worry allows for less imagery, Borkovec and Inz realized that this same outcome could alternately or additionally ref lect the motivated use of worrisome thought in order to avoid negative images. Further hints that this might be the case have come from Stöber’s (1998; Stöber, Tepperwien, & Staak, 2000) work on the abstract nature of worry. Stöber and colleagues have found that nonworry topics are associated with more imagery than are worry topics, and that increased worry about any given topic is associated with lessened imagery. They have also found that whatever imagery does emerge from worry topics is objectively rated to be less concrete. Paivio (1986) has shown that forming images from abstract (as opposed to concrete) words and sentences requires a greater amount of time and produces images that are subsequently rated as more abstract and less vivid. Given what is known about the nature of worry, and given dual-coding theory, the findings of Stöber’s group make sense. They suggest that the conceptual, verbal nature of worry largely limits conscious accessibility to parallel-processed images, and that even when occasional catastrophic images do occur in awareness, they are likely to be less vivid, intrusive, attention-grabbing, and emotionally distressing. Again, such conditions suggest the possibility of motivated cognitive avoidance. If motivated cognitive avoidance is in fact involved in the abovedescribed effects, what might be the imagery mechanisms that are susceptible to worry’s mitigating inf luence? According to Kosslyn (1983), imagery involves both a retrieval mechanism and a refresher mechanism. Once an image is formed by retrieval from memory, a separate refreshing pro-

Avoidance Theory of Worry

83

cess is needed to maintain it. Worry probably primes the involuntary retrieval of periodic aversive images, but at the same time serves as a means of cognitive avoidance in response to them. Kosslyn would be likely to place the point of worry impact at the level of the refreshing mechanism. Shifting to worry upon each image occurrence would interfere with the refreshment of that image, and the image would fade away. During a worry episode, the spread of associations continues along the verbal/linguistic pathway until another catastrophic image is encountered and the same process occurs again. Such a process would ref lect the classic operation of Mowrer’s two-stage anxiety theory at the cognitive level.

The Suppression of Somatic Anxiety Obviously, we believe that the predominance of thought in the worry process is crucial for understanding its nature and functions, and that the distinction between thought and imagery is fundamental to that understanding. Psychophysiological research on the differences between these two types of cognitive activity has contributed significantly to why we believe this to be true. Specifically, this literature has provided an answer to how worrisome thought might be used as an avoidance response—that is, as an immediate coping response to the anxiety generated by the perception of threat, and not solely as a cognitive attempt to solve the problem of perceived future danger. The answer resides in the fact that verbal articulation of fear material elicits little cardiovascular response, whereas imagination of a scene that represents the same material evokes considerable heart rate reaction (Vrana, Cuthbert, & Lang, 1986). Applying this fact to what is known about worry, one can speculate that when aversive images occur in the process of worry and/or when threatening environmental stimuli are detected, the shifting of attention to worrisome thinking upon each occurrence or detection results in the escape from or avoidance of the somatic element of the fear response to the aversive stimuli; internal fear cues would be muted, and part of Foa and Kozak’s (1986) requirements for extinction are left unfulfilled. In this way, worry may be an attempt to suppress both the aversive images (e.g., via reduced concreteness) and the unpleasant somatic anxiety (Mowrer’s “bad states”) otherwise elicited by them. Such a process would need to continue throughout a worry episode. Thinking about worrisome concerns would necessarily prime catastrophic images related to those concerns, and thus tend to increase the frequency of such images (images and thought occur simultaneously and integratively in dual-coding theory; Paivio & Marschark, 1991). Upon each image occurrence, however, there is functional reason to shift immediately to thought activity in order to reduce the physiological/affective distress elicited by the images. Thus worry is hypothetically motivated both by the need to remove perceived threat

84

THEORETICAL AND EMPIRICAL APPROACHES

(i.e., to find a solution to the problem) and by its suppression of certain (somatic) aspects of anxious experience. Significantly, empirical research has documented that the presence of a worrisome state in fact suppresses an important type of somatic response to feared images—specifically, sympathetic (fight-or-f light) activation. Remarkably, this suppression has been found both for sympathetic activation that would otherwise occur in response to feared material and for sympathetic activation during the worry process itself.

Suppression in Response to Fear Material Over 20 years ago, Tucker and Newman (1981) found that individuals spontaneously engage in verbalization as a strategy to achieve disengagement and emotional control while decreasing their sympathetic arousal to distressing, aversive information. Borkovec and Hu (1990) discovered several years later that speech-anxious participants who engaged in worrisome thought prior to each of several presentations of a phobic image showed very little cardiovascular response to those images, despite reporting heightened subjective anxiety. In contrast, participants who engaged in either relaxed thinking or neutral thinking showed strong heart rate reaction to the images (with declining reaction over repeated images in the relaxed-thinking group). In a partial replication of this effect, Borkovec, Lyonfields, Wiser, and Diehl (1993) found that this suppression of somatic response was linked to the amount of worrisome thinking that occurred and not merely the amount of thinking in general. Peasley-Miklus and Vrana (2000) also found lessened heart rate response to feared imagery following worrisome thinking than after relaxed thinking when the thinking periods were used as the baseline for measuring the imagery response (although the effect did not occur if the baseline was taken from prethought periods). All three investigations indicated that the actual somatic reactions to images containing feared material are muted by preceding worry. Thus, throughout the process, worry periodically and immediately produces some removal of Mowrer’s “bad states”; this is supportive of the notion that worry acts as a negatively reinforced avoidance activity. By so doing, worry will simultaneously preclude exposure to certain aspects of internal fear cues—a fact that will play a significant role when we discuss the implications of worry for the maintenance of anxiety, as well as for the development of therapeutic interventions for GAD. One recent investigation (Hazlett-Stevens & Borkovec, 2001) qualifies the generality of this suppression effect. Preceding worry did not lessen physiological responding during repeated in vivo exposures to public speaking among speech-anxious participants. It did result in the greatest subjective anxiety experienced during the exposures compared to preced-

Avoidance Theory of Worry

85

ing neutral and relaxation conditions (in replication of the Borkovec & Hu [1990] investigation). Hazlett-Stevens and Borkovec argued that public speaking inherently generates a high degree of physiological activation, given its complex cognitive–motor demands, and that this may have overwhelmed any effects of the worrisome cognitive state. Despite the physiological evidence that the entire fear structure had been accessed in all conditions, they also argued that the greater subjective fear responses to the repeated speeches shown in the worry condition suggested a maintenance of anxious meaning and thus a preclusion of extinction, due instead to the possible absence of the second requirement in Foa and Kozak’s (1986) model (i.e., the processing of corrective information). If one worries before and after exposure to a feared event, internal or external information that nothing bad is happening or has happened is less likely to be processed. Although investigation of the impact of worry on emotional processing during in vivo exposure to other anxiety-provoking situations is warranted, at present we can confidently conclude only that worry suppresses somatic responses to imaginal material.

Suppression during Worrisome Episodes Although worry’s specific inhibition of the sympathetic activation during images of feared material is a significant piece in the puzzle of worry mechanisms, it turns out that both GAD and worry are associated with a general suppression of sympathetic activity as well. The only peripheral psychophysiological measure shown to be elevated in GAD, both at rest and in response to challenge, has been muscle tension. Otherwise, despite their connection to anxiety, both the condition of GAD and the state of worry are not characterized very often by the sympathetic activation so prevalent in other anxiety disorders. Instead, they are associated with reduced autonomic variability, indicative of sympathetic inhibition and/or deficient parasympathetic tone (Hoehn-Saric & McLeod, 1988; HoehnSaric, McLeod, & Zimmerli, 1989; Lyonfields, Borkovec, & Thayer, 1995; Thayer, Friedman, & Borkovec, 1996). Clients with GAD even show this specific type of rigid autonomic functioning during resting states. Moreover, the induction of worry, even among those who do not habitually worry, causes similar effects during its occurrence. It is significant that several experimental tasks known to decrease autonomic variability (mental arithmetic, recall of a past aversive event, threat of shock, and isometric grip; Grossman, Stemmler, & Meinhardt, 1990; Grossman & Svebak, 1987) involve processes that are decidedly and tonically characteristic of clients with GAD. Paralleling these physiological data, Marten and colleagues (1993) found that only symptoms significantly mediated by the central nervous system (restlessness or feeling keyed up or on edge, disturbed sleep, muscle tension, irritability, being easily fatigued, and diffi-

86

THEORETICAL AND EMPIRICAL APPROACHES

culty concentrating) were frequently and reliably reported by clients with GAD, whereas symptoms mediated by the autonomic nervous system were less common. As a result of this empirical observation, the former symptoms were retained and the latter were deleted from the associated symptom checklist for GAD in DSM-IV (APA, 1994). It is also significant that reduced parasympathetic tone, characteristic of GAD, is implicated in attentional control mechanisms (see Porges, 1992)—possibly providing a basis for the attentional difficulties and biases documented in this disorder (see MacLeod & Rutherford, Chapter 5). What mechanisms might be responsible for worry’s specific suppressing effects on somatic reactions to phobic imagery and/or its effects on sympathetic activity in general? Although the answer remains speculative, a few possibilities appear reasonable. First, because worry absorbs attentional and other cognitive resources, it may be difficult to shift attention away from the worry process to some other stimulus (such as a phobic image). Second, Gray’s (1982) neurophysiological theory of anxiety suggests that this emotion will occur when information received from the environment is aversive and/or when it differs from the information expected. Notice that in the phobic imagery studies, a preceding period of worrying creates an internal state in which the external information matches significant elements of the content of the worrying. Under this circumstance, there is a reduction in the degree of mismatch between information received and information expected, removing one of the anxiety-eliciting conditions in Gray’s theory. Third, if worry is primarily thought, and if thought is less closely tied to efferent command than is imagery (Lang, 1985) or real-life events, then some sympathetic suppression, both in general and in response to threat, derives from the very (thinking) nature of worry. Whatever mechanisms are at work to explain why thought is not strongly connected to efferent command (and therefore to affect, physiology, and behavior) should reside within the neuroanatomy of cortical and subcortical functioning. But it is at least clear that evolution was wise to create a conceptual system that is, on the one hand, free from immediate behavioral expression and the potential punishing consequences of such expressions—and, on the other hand, free to play and experiment mentally with possible choices of action before implementing an action. Fourth, worry may involve semantic satiation. Interviews of clients with GAD and introspective observations both suggest that worrying about a concern tends to be repetitive (i.e., mentally saying basically the same things over and over). Empirically, content analysis of stream-of-consciousness reports has shown that worry is characterized by significant decreases in topical shifts (Molina, Borkovec, Peasley, & Person, 1998). The cognitive psychology literature has demonstrated that the repetition of a word leads to a temporary weakening of the links to the rest of its associative network of

Avoidance Theory of Worry

87

meaning (e.g., Smith, 1984). By implication, if worrying contains the repetition of crucial, worry-related words, this repetition will result in lowered access to other elements of meaning (especially the affective and other response elements) contained within their network. Finally, the combination of earlier-reviewed empirical features of GAD and worry also provides hints about other mechanisms that may interconnect some of those features. Clients with GAD are detecting threat frequently. Threat perception ordinarily (and evolutionarily, for the sake of survival) elicits fight-or-f light sympathetic activation to facilitate avoidance. There is rarely an overt behavioral avoidance response available to clients with GAD to remove the threat, and so they naturally turn to cognitive attempts to solve the problem. The problem, however, is frequently insoluble, because (1) the threat cues perceived exist solely in the mind; (2) those cues refer to possible bad events in the future, which do not actually exist in the present moment; and (3) the vast majority of bad events that are feared are not going to happen anyway (as we will document later). Under these circumstances (a fight-or-f light response with no place to go), sympathetic activation is not useful or adaptive and so is suppressed. Animal analogues exist wherein precisely these kinds of circumstances and effects occur. When some organisms are trapped in a threatening environment and escape is not possible, they freeze (increased muscle tone, reduced autonomic variability, and sympathetic inhibition).

Suppression of Affect in General Consideration of the above-described findings about the nature of worry and GAD leads to the speculation that perhaps clients with GAD are less in tune with affective experience in general, and not merely with the somatic aspects of anxious responding. The only empirical evidence indirectly related to this possibility comes from studies finding that college students meeting criteria for GAD reported greater difficulty in identifying and describing their emotions on the Toronto Alexithymia Scale than did nonanxious controls (Abel, 1994; Yamas, Hazlett-Stevens, & Borkovec, 1997). Either persons with GAD may be actively avoiding emotion, or their constant focus of attention on thought activity may make them less aware of affective f luctuations throughout the day. Research focused on more valid assessments of general emotional experience in GAD is required to verify this speculation, but if it turns out to be correct, then future therapy developments for GAD could usefully incorporate measures to increase clients’ processing of emotion in general. Although worry has been demonstrated to suppress immediate somatic anxiety responses to threatening material and thus appears to serve as a negatively reinforced cognitive avoidance activity, it is also important

88

THEORETICAL AND EMPIRICAL APPROACHES

to examine the evidence for our earlier logical conclusion that worry functions as an attempt to solve the problem of possible future bad environmental events and thus acts as an effort to avoid distal dangers as well.

Worrying to Avoid Future Bad Events We have begun this chapter with the self-evident logic behind the proposition that worry functions to anticipate and avoid bad outcomes in the future. Empirical evidence indicates that people do at least believe that worry is an effective method of preventing disaster (see also Freeston, Rheaume, Letarte, Dugas, & Ladouceur, 1994, and Wells, Chapter 7, this volume, for a comprehensive discussion of both positive and negative beliefs about worry). Several years ago, Borkovec and his colleagues asked clients with GAD undergoing the Penn State treatment protocols (to be described later) what benefits they saw to worrying. Six common themes emerged. Worry was believed to be a way to (1) motivate oneself to get things done, (2) determine how to avoid or prevent bad events, (3) prepare for the worst, (4) problem-solve, (5) superstitiously lessen the likelihood of bad events, and (6) distract oneself from even more emotional topics. Each of these potential benefits of worrying functionally contains a negative reinforcement contingency (see Borkovec, Hazlett-Stevens, & Diaz, 1999, for a more detailed description of this for each belief). Results from a questionnaire constructed from these items and administered to college students who met or did not meet GAD criteria showed that all six reasons to worry were widely endorsed by both groups, with motivation, avoidance, and preparation being the most highly rated (Borkovec & Roemer, 1995). In addition, participants with GAD scored significantly higher on the item involving distraction from more emotional topics in two separate studies. Although most of the items ref lect both psychodynamic and learning theory views of the neurotic paradox, this latter item, which distinguished GAD from nonanxious groups, is perhaps the quintessential example: Humans may create distress for themselves in order to preclude even greater subsequent pain. It may be the case that merely believing that a negative reinforcement contingency exists is sufficient for an activity like worry actually to be maintained. However, some of the empirical research mentioned earlier and additional research described below provide evidence that a functional basis may indeed exist for some (perhaps all) of the positive beliefs about worry listed above. Negative reinforcement requires that a behavior be followed by the removal of a punishing state. The punishing state includes both the bad environmental events and the feelings of fear themselves. In Mowrer’s (1947) terms, the rat’s avoidance response terminates both the internal (conditioned response) fear cues and the possibility of being exposed to punishment (the aversive unconditioned stimulus). If we

Avoidance Theory of Worry

89

found evidence that worry is associated with either or both of these events, then we would have evidence consistent with the hypothesis that worry is potentially being negatively reinforced. Given that worry immediately mitigates somatic reactions to phobic images (a direct demonstration of its ability to reduce feelings of fear and create Mowrer’s “good” state), it may well provide sufficient historical experience to support a belief that worry does indeed prepare one for the worst. If, in addition, one has multiple experiences of worrying about potential bad events, and the worry process itself is associated with lessened somatic anxiety, then one might reasonably conclude that worry is an effective way to prepare oneself for catastrophe. In terms of negative reinforcement from the avoidance of more temporally distant punishing environmental events (solving the problem presented by perceived threat), there is clear evidence that the vast majority of things about which people worry never happen (Borkovec et al., 1999). College students with and without GAD, as well as clients with GAD from our therapy protocols, have kept daily diaries over 2 or more weeks in which they noted every topic about which they worried and then rated the eventual environmental outcomes once they happened. The rating for each outcome was made on the basis of whether the outcome turned out as bad as feared or better or worse than feared. Seventy percent of the students’ worries turned out better than they had feared. This was true regardless of GAD status, although the participants with GAD spent much more time worrying during a typical day. For the clients, 85% of their worries had positive outcomes. Because clients with GAD typically fear not being able to cope with negative events, the clients in this research also rated how well they coped with those bad outcomes that did happen. The results indicated better-than-expected coping in 79% of those cases. What these data importantly indicate is that worry is typically followed by the nonoccurrence of the feared event. This was especially the case for clients with GAD, and the nonoccurrences included both the feared environmental catastrophes and the feared inability to cope. Under these schedules of negative reinforcement, the chronic worry seen in GAD is both strengthened by frequent negative reinforcement and becomes resistant to extinction, due to the type of schedule of reinforcement that the world provides. Given these contingencies, a direct functional basis for two of the positive beliefs about worry can be seen—namely, that worry can help one to find effective ways to avoid or prevent bad events, and that worry somehow superstitiously lessens the likelihood of such events. Although the possibility has not yet been empirically tested, it is likely that self-monitoring of outcomes related to two of the other beliefs (problem solving and motivation) would show similar results indicating superstitious reinforcement. Worry is necessary neither for engaging in problem solving (in-

90

THEORETICAL AND EMPIRICAL APPROACHES

deed, anxiety generally interferes with complex task performances) nor for motivating adaptive behavior (positive contingencies could be adaptively substituted to achieve the same goal). In either case, however, the outcomes are quite likely to turn out better than feared, regardless of whether the person chooses positive or aversive self-control strategies, and will thus similarly ref lect the operation of a variable schedule of superstitious reinforcement.

Worrying to Avoid More Emotional Topics No direct evidence exists as yet for the possibility that worry might function as a cognitive means of avoiding more emotional topics. If indeed such topics do exist, and if worry does in fact function to avoid exposure to them, however, then there are intriguing possibilities for what those more emotional topics might be. These possibilities come in two forms with regard to likely associative network organization.

Topic-Related Underlying Fears In the vertical form of network organization, it may be that the conscious content of a worry typically involves only the most superficial elements of its worrisome fear structure and does not access deeper representations of the threat underlying those elements. Under this circumstance, superficial worrisome thoughts are distractions from, and therefore ways of avoiding, thoughts about the more anxiety-provoking levels of content. The use of decatastrophizing methods in cognitive therapy assumes that this is the case; such methods are used to access the clients’ deeper underlying fears, ref lected by the more distal anxious thoughts typically reported by clients initially. One study is potentially supportive of the possibility that worry is avoidant of deeper threat (Vasey & Borkovec, 1992). “Worriers” and “nonworriers” were asked to indicate what would be bad or fearful about each step in the process of catastrophizing their most worrisome topic. Worriers generated a significantly greater number of steps, and most importantly, they showed increasing levels of discomfort over the consecutive steps; nonworriers maintained low levels of discomfort throughout the catastrophizing sequence. Thus some of the conditions necessary to support the hypothesis that chronic worriers worry at a superficial level in order to avoid more anxiety-provoking material were present—namely, a rich associative network of worry-related steps, with increasing discomfort during the accessing of those steps. In the horizontal form of network organization, worrying about a particular topic may function to distract a person from completely different content—content that is more anxiety-provoking in nature. What such content might be has not been empirically established, but research on

Avoidance Theory of Worry

91

GAD has determined at least three domains in which significantly negative historical or current events may contribute to negative emotional experiences about which the person would rather not think.

Past Trauma Theories of GAD have proposed that stressful life events play a causal role in two of the central characteristics of the disorder—namely, chronic anxious apprehension (Barlow, 2002) and excessive perceptions of the world as a dangerous place (Beck & Emery with Greenberg, 1985). Indeed, Beck and colleagues (1985) suggested that exposure to a single potentially traumatic life event might be sufficient for the development of GAD. Furthermore, traumatic life events occurring after the onset of GAD would reasonably contribute to its maintenance or exacerbation, given that such events would be likely to reinforce an individual’s perceptions of the world as threatening and/or perceptions that he or she may not be able to cope. If past trauma is frequent among people suffering from GAD, then it may represent the “more emotional topic” from which their daily worry and its content are a distraction. In both undergraduate and client samples, individuals meeting criteria for GAD have been found to report a significantly greater frequency of past traumatic events than participants without GAD (Roemer, Molina, Litz, & Borkovec, 1997). Similarly, a GAD diagnosis is associated with a higher frequency of at least one unexpected, major negative life stressor (Blazer, Hughes, & George, 1987). Of course, a reported history of trauma is common as well in persons with other psychiatric disorders, including other anxiety disorders (Fierman, Hunt, Pratt, & Warshaw, 1993; Herman, Perry, & van der Kolk, 1989; Jordan et al., 1991). Thus a history of trauma may simply make indiv iduals more susceptible to psychopathology in general, as opposed to GAD in particular. Some empirical research offers hints of specificity, however. Zuellig, Newman, Alcaine, and Behar (1999) conducted childhood diagnostic interviews with three groups of college students: those who met criteria for adult GAD, those who met criteria for adult panic disorder without GAD, and those who met no adult diagnostic criteria. They found significantly higher rates of retrospectively reported childhood posttraumatic stress disorder in the group with GAD. Similarly, clients diagnosed with GAD have reported experiencing the death of a parent before the age of 16 significantly more often than clients with panic disorder (Torgersen, 1986). Although further comparisons with other anxiety disorder groups would be important, extant data do suggest that individuals with GAD may indeed have more traumatic histories than do individuals with at least one other anxiety disorder. Caution is advised in interpreting these data, given that retrospective reports can in general be unreliable; mood-congruent

92

THEORETICAL AND EMPIRICAL APPROACHES

memory may play a role in selective recall and/or interpretation of past events by a client with GAD; only 50% of the group with GAD in the Roemer, Molina, Litz, and Borkovec (1997) study reported traumas (indicating the likelihood that other historical factors must be playing an etiological role); and no investigation to date has explored whether the reported traumatic experiences occurred prior to or after the onset of GAD. If the past experiences of some people with GAD do indeed contain more traumatic events than is true for other individuals, then investigating the relationship between the content of their traumas and the content of their worry topics might reveal evidence about whether or not these individuals might be using worries about current concerns to avoid memories of those past experiences. In the Roemer, Molina, Litz, and Borkovec (1997) investigation, groups with GAD reported more frequent occurrences of such events as death, illness, injury, and assault. Importantly, however, these are precisely the topics about which people with GAD worry the least (Roemer, Molina, & Borkovec, 1997). So, although these individuals report having experienced these traumatic occurrences, they now fail to worry about those topics that are most directly relevant to the past traumas that they have experienced. This important discrepancy leads us to consider two possibilities: Either they have habituated to cognitive activity dealing with traumatic material about physical threats, or they are avoiding memories about these more emotionally distressing topics by worrying about other things.

Early Childhood Relationships Extremely stressful or life-threatening events are of course not the only historical aversive experiences that people might wish to avoid in their daily recollections. Recent investigations into what individuals with GAD remember from childhood relationships with their primary caregivers provide another clue about what the avoided content of their “more emotional topics” might be. Bowlby (1982) hypothesized that insecure attachments in early childhood may lead to anxiety later in life. Attachment, he theorized, is a child’s attempt to seek safety and comfort whenever threat is perceived in the environment. If the child repeatedly experiences the unavailability of an attachment figure, the child may come to perceive the world as a dangerous, frightening place. The result may be a tendency to overestimate the likelihood and severity of feared events, as well as to underestimate resources available to cope with those threats. Such perceptions are remarkably similar to those of adult clients with GAD. Clients with GAD (Cassidy, 1995) as well as undergraduates meeting criteria for GAD (Cassidy, 1995; Schut et al., 1997), have completed the Inventory of Adult Attachment (Lichtenstein & Cassidy, 1991), a question-

Avoidance Theory of Worry

93

naire designed to assess reports of attachment-related memories from childhood. In both investigations, the groups with GAD reported more role-reversed/enmeshed relationships: As children, they had to take care of their primary caregivers (instead of being taken care of). Thus they had to anticipate dangers not only for themselves but for those primary caregivers as well. This provides an understandable, early childhood basis for the view of the world in adulthood GAD—that is, as a potentially threatening place where the person’s ability to cope is limited and threat must be constantly anticipated. Also in both studies, the groups with GAD indicated greater current levels of unresolved anger toward and vulnerability to their childhood primary caregivers. Although the groups with GAD in the Cassidy (1995) and Schut and colleagues (1997) studies reported the existence of these role-reversed/ enmeshed relationships with residual anger and vulnerability, they also reported having greater difficulty remembering their childhood relationships. Although alternative hypotheses exist (e.g., chronic, constant thought activity may simply make recall of distant past events more difficult), it is also possible that this report indicates an avoidance of negative attachment-related memories. As would be the case for the many other instances of negative emotional experience in the daily life of a client with GAD, worry may be avoidant by filling the mind with superficial conceptual activity when it might otherwise shift to thoughts of a distressing and anger-inducing childhood, while simultaneously inhibiting the imagery and somatic activation potentially accompanying those memories.

Current Interpersonal Relationships Perhaps past trauma or negative childhoods are not the sources of difficulty that lead individuals with GAD to claim that their worrying distracts them from more emotional topics. It may be the case that the hypothetical horizontal distraction has to do with present life concerns that are simply too difficult to allow very frequently into awareness. Although clients with GAD worry about a multitude of topics, interpersonal situations are particular causes of concern and anxiety for them. Social phobia is the most frequent comorbid diagnosis for GAD (Barlow, 1988; Brown & Barlow, 1992); trait worry correlates more highly with social fears than it does with nonsocial fears (Borkovec, Robinson, et al., 1983); and content analysis of worry topics reveals that interpersonal concerns are more frequent than any other topical area (Roemer, Molina, & Borkovec, 1997). More exciting has been the identification of specific types of current interpersonal concerns among clients with GAD (Pincus & Borkovec, 1994), as determined by the Inventory of Interpersonal Problems (Horowitz, Rosenberg, & Bartholomew, 1993). Not only did the group with GAD show greater levels of distress and rigidity in their inter-

94

THEORETICAL AND EMPIRICAL APPROACHES

personal relationships than did nonanxious controls, but cluster analysis revealed that the largest grouping (62.1% of clients with GAD) involved an interpersonal style characterized by being “overly nurturant and intrusive” in their relationships (the other two, smaller clusters were “cold/vindictive” and “socially avoidant/nonassertive”). What is exciting about this outcome is that it connects conceptually so well with the childhood attachment findings. It appears that the majority of the clients with GAD learned in childhood to take care of others in order to receive love and nurturance, and that they continue to do this in their adulthood relationships, possibly for the same reason. Worry may thus function as a method for anticipating the needs of significant others in an attempt to satisfy one’s most important interpersonal needs and/or as a distraction from the realization that those needs are indeed not being met. Although we have speculated earlier that people with GAD may experience lowered affect in general, they may, on the other hand, develop a particular sensitivity to other people’s emotional experience. This would make sense, because role reversal in childhood as well as nurturant roles in current relationships would require that they learn to detect the feeling states of significant others in their lives. Indeed, participants meeting GAD criteria have been found to report significantly greater empathy with others, and especially greater ability to feel the pain of others, than nonanxious participants report (Peasley, Molina, & Borkovec, 1994).

IMPLICATIONS OF WORRY AS A NEGATIVELY REINFORCED COGNITIVE AVOIDANCE RESPONSE The review above empirically documents several of the immediate consequences of worry. The worry process, however, has potentially broader implications, described below.

Emotional Processing The most significant implication of worry as cognitive avoidance is, of course, the fact that it probably precludes emotional processing and thus maintains anxious meanings. In Foa and Kozak’s (1986) model, the entire fear structure must be accessed if extinction is to take place. This structure contains stimulus, response, and meaning propositions (Lang, 1985). Failure to access any of these elements implies that extinction will be retarded or nonexistent during either in vivo or imaginal exposures to disorder-relevant stimuli, and regardless of whether these exposures are taking place in daily life or in systematic therapeutic applications. Overt behavioral avoidance obviously precludes exposure. Analogously, if worry is functioning to help a person escape or avoid each periodic occurrence of

Avoidance Theory of Worry

95

catastrophic images, it minimizes exposure to the imaginally represented stimulus elements that represent the threat. If worry suppresses somatic reactions to such feared material, it reduces exposure to crucial response elements (in Mowrer’s terms, the conditioned fear responses) of its fear structure. If worry is superficial in either a horizontal or a vertical sense, it prevents exposure to the stimulus, response, and meaning elements of the underlying fear. The research reviewed throughout this chapter supports the conclusion that worry does indeed have these characteristics and consequences. What is even more remarkable is that separate research has shown that worry can interfere with a dramatically different type of emotional processing. If one imagines an unconditioned stimulus (UCS) previously employed in an aversive conditioning task, this UCS rehearsal will enhance subsequent fear responses to a previously established conditioned stimulus (Jones & Davey, 1990). If one worries prior to the UCS rehearsal, however, no enhancement of fear occurs, whereas either prior neutral thinking or somatically anxious thinking does strengthen the conditioned response (Davey & Matchett, 1994). This is remarkable because it indicates that worry reduces extinction effects during repeated exposure tasks, but at the same time mitigates the strengthening of anxiety during UCS rehearsal tasks. Because these are diametrically opposite outcomes whose common ingredient is emotional processing, the impact of worry on both processes suggests that its pathway of effect is through interference with that processing. A separate literature shows that worry after exposure to a stressor also mitigates adaptive emotional processing of the distressing material. Both Butler, Wells, and Dewick (1995) and Wells and Papageorgiou (1995) found that intrusive images during the 3 days subsequent to the viewing of a stressful film were greater among participants who had worried for 4 minutes about the just-seen film in comparison to control conditions. In the former study, worry also had an immediate effect on reducing subjective distress about the film. The fact that this effect was not replicated in the latter study suggests that worry sometimes does and sometimes does not produce some immediate relief from exposure to distressing material. When it does, another example of potential negative reinforcement of worry is provided.

Strengthening of Anxious Meanings Worry may also have consequences beyond mere prevention of the otherwise naturally occurring extinction effects of repeated exposure. Worry may, by its process or content, actually increase anxious meanings as well. Although worry reduces some aspects of fear responding, it still generates aversive emotional states and includes periodic aversive images. In the

96

THEORETICAL AND EMPIRICAL APPROACHES

process of worry, worrisome thinking is not only occurring in constant response to periodic aversive images, but is also followed by such images. In such a sequence, conditions exist for evaluative conditioning to take place, with anxiety-provoking images serving as UCSs. We already know that images can function as UCSs (Davey & Matchett, 1994; Jones & Davey, 1990). A more recent study has shown that even worry-related words can serve that function: Clients with GAD developed a conditioned orienting response to a neutral cue paired with threat words, whereas nonanxious control participants did not (Thayer, Friedman, Borkovec, Johnsen, & Molina, 2000). This finding also describes one possible mechanism for the growth of hypervigilance in GAD. Moreover, only the clients with GAD showed cardiovascular accelerations to the threat word presentations. This finding suggests the possibility of motivated inattention (cognitive avoidance) (Jennings, 1986) and provides psychophysiological evidence for Mathews’s (1990) view that worry sequentially involves ongoing selective vigilance to threat followed by evasion of that threat. Finally, the very act of avoiding something (overtly or covertly) may, from the James–Lange perspective on the nature of emotion (see Fehr & Stern, 1970), strengthens the anxious meanings associated with stimuli so avoided (“If I am worrying about it, the topic must contain danger”). An analogous process has been shown in the thought suppression literature, wherein suppression even of a neutral thought leads to an increase in the anxiety-provoking value of that suppressed thought (Roemer & Borkovec, 1994).

Relevance to Other Anxiety Disorders Based on early age of onset, chronicity, and lack of treatment responsiveness, Brown, Barlow, and Liebowitz (1994) have argued that GAD may be the basic anxiety disorder out of which other anxiety and mood disorders emerge. Furthermore, although worry is the central defining feature of GAD, it is also pervasive in these other disorders (Barlow, 1988), and GAD is one of their most frequent comorbid conditions (Brown & Barlow, 1992). Given these observations, it follows that everything we know about the nature and functions of worry becomes potentially relevant to the understanding and treatment of any disorder of the emotions. If worry precludes emotional processing, and if other emotional disturbances require such processing for their amelioration, then worry may be a factor in their maintenance and/or therapeutic attempts to reduce them. Surprisingly, however, little systematic research on the role of worry in other disorders has yet been conducted; such research remains as a potentially crucial area for further empirical pursuit. One investigation that underscores the potential importance of this domain discovered that suc-

Avoidance Theory of Worry

97

cessful treatment of principal GAD eliminates nearly all comorbid conditions by follow-up assessment (Borkovec, Abel, & Newman, 1995).

Relevance to Physical Health The habitual, rigid nature of the cognitive, physiological, and interpersonal systems found to exist in GAD may well have health implications beyond psychological dysfunction. This appears to be particularly likely because of their distinctive psychophysiologies, characterized by parasympathetic deficiency. Evidence strongly suggests that reduced physiological variability (especially in the cardiovascular system) is implicated in numerous physical pathologies, such as sudden cardiac death, hypertension, ventricular fibrillation, and coronary atherosclerosis (Goldberger, 1992). It is not surprising, then, that a 20-year longitudinal study found that worry predicts an increased incidence of coronary heart disease (Kubzansky et al., 1997); that worry affects the immune response to phobic stimuli (Segerstrom, Glover, Craske, & Fahey, 1999); and that GAD is associated with significant utilization of family doctors and other (nonpsychiatric) medical specialists (Greenberg et al., 1999), as well as with significant social and occupational impairment (Roy-Byrne, 1996).

Absence of Present-Moment Focus of Attention The most poignant feature of a life spent lost in one’s thoughts, especially thoughts that constantly create negative experiences based upon the illusion of nonexistent futures, is that such a way of being disconnects the person from present-moment experiences—which are, in the final analysis, the only reality that is available to us. It is simply a matter of attentional resources: The more attention is allocated to an illusory inner world, the less attention can be devoted to what is right in front of us. Primary emotional experiences (as opposed to secondary or instrumental emotions; see Greenberg & Safran, 1987) in response to immediately available events, the information contained in the world from which we can learn and grow, connection to the people in our lives—these things that give us aliveness and adaptiveness are compromised in their positive inf luence if they are not attended to and processed. Perhaps this is the characteristic of GAD that is most in need of change.

THE SEARCH FOR EFFECTIVE THERAPIES FOR WORRY AND GAD The Traditional Cognitive-Behavioral Approach to GAD As mentioned before, GAD did not lend itself well to the exposure models and techniques of early behavior therapy, given its diffuse nature, lack

98

THEORETICAL AND EMPIRICAL APPROACHES

of identifiable and circumscribed anxiety elicitors, and consequential absence of clear avoidance behaviors. The growing use of cognitive therapy with anxiety disorders, and a shift to coping-oriented rather than masteryoriented therapeutic approaches, provided at least a beginning point for the development of psychological interventions for this previously neglected disorder. By the early 1980s, a package of cognitive-behavioral therapy (CBT) techniques had been created and applied to the treatment of GAD (see Leahy, Chapter 11). These techniques typically involved three primary elements. First, the clients were taught relaxation methods as generalizable skills; emphasis was placed on deployment of relaxation strategies throughout the day and at any time when anxiety or worry might be noticed (e.g., Öst’s [1987] applied relaxation method). Second, because GAD involves frequent detection and/or interpretation of threat, cognitive therapy, often based on Beck and colleagues’ (1985) technique descriptions, was employed to help the clients generate alternative ways of perceiving and believing that were less anxiety-provoking. Third, some element of rehearsal of these relaxation and cognitive skills was often provided, in order to build up their habit strength and to make it more likely that clients would remember to make use of these coping strategies throughout their daily life and in response to stressors and stress (e.g., Suinn & Richardson’s [1971] anxiety management training and Goldfried’s [1971] self-control desensitization). In each instance, clients repeatedly imagined incipient anxiety or worry and their associated cues, and then imagined applying their newly learned coping responses within those situations. Although controlled clinical trials conducted on various versions of this CBT package have documented its efficacy, reviews of this literature (Borkovec & Ruscio, 2001; Borkovec & Whisman, 1996; Chambless & Gillis, 1993; see also Gould, Safren, Washington, & Otto, Chapter 10, this volume) have also indicated that only about half of the clients so treated achieve high end-state functioning at follow-up assessments. It is clear that further improvements in therapy for GAD are needed.

The Current CBT Approach at the Penn State GAD Project Our current CBT approach being used in our GAD efficacy research at Penn State is grounded in the techniques described above, but we have modified or expanded those interventions based on our growing knowledge about the nature and functions of worry and GAD. Below we brief ly describe some salient examples of these treatment developments.

Response Prevention To confront their fears, people need to prevent their avoidance responses.

Avoidance Theory of Worry

99

But how to stop uncontrollable worry is precisely the problem that the client with GAD faces. Its uncontrollable nature is likely due to the fact that worry is a strongly reinforced habit, and as long as the perception of threat remains and no behavioral response provides a solution to the perceived threats, worry will continue. We have found one technique that holds some partial promise for reducing the amount of worrying that is taking place. Because people can worry at any time and in any place as long as the environment is not demanding their attention, worry as an operant activity becomes associated with, and thereby is triggered by, many times and places; it is under poor discriminative control. Although telling clients with GAD to stop worrying and be happy will not be effective, they can comply successfully with a request to gradually restrict the time and place in which their worrying is occurring in order to achieve some degree of stimulus control of the activity. In its empirically validated version (Borkovec, Wilkinson, Folensbee, & Lerman, 1983), this stimulus control program instructs people to establish a worry period at the same time and place each day; to monitor their worrying during the day and to learn to recognize its initiation; and to postpone the worrying upon detection to the worry period, reminding themselves that they will have the opportunity to devote their full energy to worrying and problem solving later, and that they can more adaptively focus their attention now on whatever task is at hand. A second version created by one of our therapists, Mary Boutselis, involves determining a daily time, place, or activity that will become a “worry-free zone,” wherein any detected worrying is merely put off until the client is outside that zone. Over sessions, the client expands the zone to include more times, places, and activities.

Worry Outcome Monitoring As we have described earlier, because many individuals with chronic worry or GAD believe that worrying serves some adaptive functions (e.g., prevention of bad events and preparation for the worst), we ask clients to (1) write down in a daily diary each worry that they notice during the day and its feared outcome; and (2) evaluate at the end of each day any outcomes relevant to a diary entry, what the actual outcome was, and how well they coped with it. This process results in their being frequently exposed to significant corrective experiences. They discover that most of the things that they worry about do not happen, and that they cope quite well with those that do. At the same time, they are building in memory a history of accurate evidence from which they can draw for making predictions in the future. All of this information contributes to a decline in their general tendency to perceive threat in their futures.

100

THEORETICAL AND EMPIRICAL APPROACHES

Present-Moment Focus of Attention One of the ultimate goals for both relaxation techniques and cognitive therapy methods is to have clients learn to live more fully in the present moment. Clients with GAD spend excessive time living in a largely negative world that does not exist. This illusion is created by their focus of attention on worrisome thoughts and images that refer to the past or to possible (though highly unlikely) futures. If they are to find relief from this incessant world, they must usefully learn to shift their attention to what is real, and the only reality that exists is in the present moment right before their eyes. It is in this present moment that adaptive information, new learning, growth, and development exist. Because anxiety is always anticipatory, by definition there can be no anxiety in the present moment. So our clients are encouraged to practice focusing their attention more and more often on the present moment (e.g., the task at hand, the sensations provided by the external world, or the information being offered by that world). Relaxation practice and applied relaxation, with their inherent request to “let go” of everything (including thoughts) and to focus only on the feelings of relaxation, provide initial experiences of pleasant present moments. Over time, clients are increasingly asked to practice opening up and attending to other present-moment experiences being offered by the world from this relaxed and tranquil place. Cognitive therapy also proceeds in stages leading to the present moment. In the first stage, we help our clients to recognize the frequent inaccuracy of their predictions about the future and to replace these predictions, as soon as they are detected, with alternative perspectives that are more accurate. In the second stage, we encourage them to notice that no prediction about a future event is ever completely accurate, and we invite them to begin to spend increasing time living an “expectancy-free” life. Clients who have begun to recognize through their self-monitoring that things usually turn out fine, and that they are indeed quite successful in coping with whatever comes down the road, can discover that living in the present without the need to think very much about the future at all can be a very pleasant way to be. The goal is to encourage them to let go of the felt need for certainty in an uncertain world and to trust themselves in the face of that reality (see Dugas, Buhr, & Ladouceur, Chapter 6, this volume, for other approaches to the intolerance of uncertainty). In the third stage, once clients are focused more on the present, we can begin to work with them on ways of constructing the meaning of that present moment. We now encourage our clients to focus on the intrinsics of whatever present moment they face. Special emphasis is placed on using perspectives that contain the clients’ values and/or can bring joy to the moment. Throughout both relaxation training and cognitive therapy, we also emphasize f lexibility, in contrast to the rigidity that is typical both in our

Avoidance Theory of Worry

101

clients’ psychophysiology and in their patterns of thought. Specifically, we teach our clients multiple methods of relaxing (progressive relaxation, meditation, slowed and paced diaphragmatic breathing, and relaxing imagery), and we make sure in cognitive therapy applications that a client and therapist generate multiple perspectives for any given event or situation. Underlying these approaches is the notion of playing and experimenting with multiple ways of being.

The Use of Imagery Because of our clients’ hypothetical escape and avoidance of negative images, we employ the typical self-control desensitization technique: Clients repeatedly imagine worrisome or other anxiety-provoking circumstances until negative emotions are being experienced, at which time they imagine relaxing themselves in the situation and shifting their perspectives to the more accurate ones developed during earlier cognitive therapy. Although this method provides some exposure to feared internal and external anxiety cues, its primary purpose is to rehearse the deployment of somatic and cognitive coping responses. Once an image is well rehearsed, we finish the imagery with the imagination of the most likely outcome, based on logic, history, and evidence. Most importantly, we ask our clients to shift to an image of the most likely outcome any time during the day when they detect a worrisome concern. In certain functional senses, imagery is reality: Images can generate the same physiological responses as real-life events do (Lang, 1985), and imagined events stored in memory can at times be confused with actual historical events (Garry, Manning, & Loftus, 1996). Such processes may be the basis for the familiar client comment that new, alternative perspectives do not initially feel as true or believable as their anxious perspectives feel. The frequent imagination of catastrophes has indeed established “evidence” in memory that such things have happened, even though they have not. Creation of a new history of evidence through imagining the most likely outcomes and verifying these outcomes in daily diaries would reasonably counteract these processes.

Identification and Targeting of Core Fears That Are Avoided If we could identify specific circumstances that create anxiety for our clients, then therapy techniques to deal with these circumstances (including customarily powerful exposure methods) could be usefully employed. As we have seen, GAD is not obviously associated with circumscribed anxiety triggers. The possibility remains, however, that core fears do exist that underlie clients’ numerous and constantly changing sources of worrisome concern. Earlier, we have reviewed research suggesting that GAD is often associated with traumatic histories, attachment-related problems in child-

102

THEORETICAL AND EMPIRICAL APPROACHES

hood, and current interpersonal problems that prevent clients from having their needs met. These may well represent the underlying source of the clients’ view that the world is a dangerous place and that they may not be able to cope. Although we have periodically made use of such themes in standard CBT applications in our previous clinical trials, we have not done so in any systematic fashion. Furthermore, the focus of CBT in our trials has largely been on current concerns, and the scientific questions of those studies have revolved around changing intrapersonal processes and precluded attempts to change interpersonal behavior. But the quest to identify functionally important, underlying core fears is likely to promote greater understanding of the etiological and maintaining conditions of GAD, and the development of therapy methods specifically targeting those fears may well provide the needed ingredients for discovering a truly effective approach for this difficult-to-treat disorder. Our current efforts to do so stem from findings concerning our clients’ characteristic interpersonal problems, and from evidence that these problems predict poor short-term and long-term outcome among clients treated solely with CBT methods (Borkovec, 1995). If insecure attachment in childhood has led to generally muted affective experience and to rigid ways of establishing relationships with others in adulthood that preclude meeting one’s most basic interpersonal needs, and if this sequence is the foundation for the origins and maintenance of GAD, then a therapy that not only provides cognitive and behavioral coping responses for anxious experience, but also accesses emotion and modifies interpersonal behavior in adaptive ways, may provide the most effective form of intervention for this disorder. We should know the answer to this question in another few years, when a clinical trial contrasting CBT with and without an interpersonal/emotional processing therapy has been completed (see Newman, Castonguay, Borkovec, & Molnar, Chapter 13, this volume).

SUMMARY Evidence suggests that worry is predominantly thought activity that focuses on attempts to avoid the occurrence of future catastrophe, and that it thereby functions as a cognitive avoidance maneuver in response to perceived threats. The worry process appears to be negatively reinforced both by immediate suppression of sympathetic reactions to anxiety-provoking material and by the more long-term nonoccurrence of predicted feared events. Worry may also be superficial in its content, ref lecting an avoidance of more distressing thoughts and emotions deriving from past traumas, negative childhood attachment experiences, or current interpersonal difficulties. As is the case for any type of avoidance response, worry precludes emotional processing and thus prevents extinction, and it may

Avoidance Theory of Worry

103

also strengthen the anxious meanings surrounding its topics and their content. In that case, worry is likely to be relevant to understanding and treating not only GAD, but any emotional disorder wherein emotional processing is important to the amelioration of the problem. Moreover, chronic worry has been associated with significant physical health problems, and its excessively negative valence and its detrimental impact on ability to live in and fully experience the present moment create an inf lexible, maladaptive, and unhappy internal world. Although various forms of CBT have been shown to be effective for GAD, half of the clients so treated fail to achieve high end-state functioning. Ideally, new therapeutic developments based on the increasing knowledge of the nature and functions of worry will lead to more effective interventions in the future. Promising examples of such developments include the monitoring of actual worry outcomes, learning increasingly to live in the present moment, imagery rehearsals of effective coping, replacing catastrophic images with images of likely outcomes, identification and therapeutic targeting of fears that potentially underlie superficial avoidant worries, and the targeting of current interpersonal difficulties.

ACKNOWLEDGMENTS Preparation of this chapter was supported in part by National Institute of Mental Health Grant No. MH-58593 to Thomas D. Borkovec.

REFERENCES Abel, J. L. (1994, November). Alexithymia in an analogue sample of generalized anxiety disorder and non-anxious matched controls. Paper presented at the annual meeting of the Association for Advancement of Behavior Therapy, San Diego, CA. American Psychiatric Association (APA). (1980). Diagnostic and statistical manual of mental disorders (3rd ed.). Washington, DC: Author. American Psychiatric Association (APA). (1987). Diagnostic and statistical manual of mental disorders (3rd ed., rev.). Washington, DC: Author. American Psychiatric Association (APA). (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author. Bandura, A. (1969). Principles of behavior modification. New York: Holt, Rinehart & Winston. Barlow, D. H. (1988). Anxiety and its disorders: The nature and treatment of anxiety and panic. New York: Guilford Press. Barlow, D. H. (2002). Anxiety and its disorders: The nature and treatment of anxiety and panic (2nd ed.). New York: Guilford Press. Beck, A. T., & Emery, G., with Greenberg, R. L. (1985). Anxiety disorders and phobias: A cognitive perspective. New York: Basic Books. Blazer, D., Hughes, D., & George, L. K. (1987). Stressful life events and the onset of a generalized anxiety syndrome. American Journal of Psychiatry, 144, 1178–1183.

104

THEORETICAL AND EMPIRICAL APPROACHES

Borkovec, T. D. (1972). Effects of expectancy on the outcome of systematic desensitization and implosive treatments for analogue anxiety. Behavior Therapy, 3, 29–40. Borkovec, T. D. (1974). Heart-rate process during systematic desensitization and implosive therapy for analogue anxiety. Behavior Therapy, 5, 636–641. Borkovec, T. D. (1979). Extensions of two-factor theory: Cognitive avoidance and autonomic perception. In N. Birbaumer & H. D. Kimmel (Eds.), Biofeedback and selfregulation (pp. 139–148). Hillsdale, NJ: Erlbaum. Borkovec, T. D. (1985). Worry: A potentially useful construct. Behaviour Research and Therapy, 23, 481–482. Borkovec, T. D. (1995, July). Implication of basic research for treatment developments in generalized anxiety disorder. Paper presented at the World Congress of Behavioural and Cognitive Therapies, Copenhagen. Borkovec, T. D., Abel, J. L., & Newman, H. (1995). The effects of therapy on comorbid conditions in generalized anxiety disorder. Journal of Consulting and Clinical Psychology, 63, 479–483. Borkovec, T. D., Hazlett-Stevens, H., & Diaz, M. L. (1999). The role of positive beliefs about worry in generalized anxiety disorder and its treatment. Clinical Psychology and Psychotherapy, 6, 126–138. Borkovec, T. D., & Hu, S. (1990). The effect of worry on cardiovascular response to phobic imagery. Behaviour Research and Therapy, 28, 69–73. Borkovec, T. D., & Inz, J. (1990). The nature of worry in generalized anxiety disorder: A predominance of thought activity. Behaviour Research and Therapy, 28, 153–158. Borkovec, T. D., Lyonfields, J. D., Wiser, S. L., & Diehl, L. (1993). The role of worrisome thinking in the suppression of cardiovascular response to phobic imagery. Behaviour Research and Therapy, 31, 321–324. Borkovec, T. D., Ray, W. J., & Stöber, J. (1998). Worry: A cognitive phenomenon intimately linked to affective, physiological, and interpersonal behavioral processes. Cognitive Therapy and Research, 22, 561–576. Borkovec, T. D., Robinson, E., Pruzinsky, T., & DePree, J. A. (1983). Preliminary exploration of worry: Some characteristics and processes. Behaviour Research and Therapy, 21, 9–16. Borkovec, T. D., & Roemer, L. (1995). Perceived functions of worry among generalized anxiety disorder subjects: Distraction from more emotional topics? Journal of Behavior Therapy and Experimental Psychiatry, 26, 25–30. Borkovec, T. D., & Ruscio, A. (2001). Psychotherapy for generalized anxiety disorder. Journal of Clinical Psychiatry, 62, 37–45. Borkovec, T. D., & Whisman, M. A. (1996). Psychological treatment for generalized anxiety disorder. In M. R. Mavissakalian & R. F. Prien (Eds.), Long-term treatments of anxiety disorders (pp. 171–199). Washington, DC: American Psychiatric Press. Borkovec, T. D., Wilkinson, L., Folensbee, R., & Lerman, C. (1983). Stimulus control applications to the treatment of worry. Behaviour Research and Therapy, 21, 247– 251. Bowlby, J. (1982). Attachment and loss: Vol. 1. Attachment (2nd ed.). New York: Basic Books. Brown, T. A., & Barlow, D. H. (1992). Comorbidity among anxiety disorders: Implications for treatment and DSM-IV. Journal of Consulting and Clinical Psychology, 60, 835–844. Brown, T. A., Barlow, D. H., & Liebowitz, M. R. (1994). The empirical basis of generalized anxiety disorder. American Journal of Psychiatry, 151, 1272–1280.

Avoidance Theory of Worry

105

Butler, G., Wells, A., & Dewick, H. (1995). Differential effects of worry and imagery after exposure to a stressful stimulus: A pilot study. Behavioural and Cognitive Psychotherapy, 23, 45–56. Carter, W., Johnson, M., & Borkovec, T. D. (1986). Worry: An electrocortical analysis. Advances in Behaviour Research and Therapy, 8, 193–204. Cassidy, J. (1995). Attachment and generalized anxiety disorder. In D. Cicchetti & S. Toth (Eds.), Rochester Symposium on Developmental Psychopathology : Vol. 6. Emotion, cognition and representation (pp. 343–370). Rochester, NY: University of Rochester Press. Chambless, D. L., & Gillis, M. M. (1993). Cognitive therapy of anxiety disorders. Journal of Consulting and Clinical Psychology, 61, 248–260. Davey, G. C. L., & Matchett, G. (1994). Unconditioned stimulus rehearsal and the retention and enhancement of differential “fear” conditioning: Effects of trait and state anxiety. Journal of Abnormal Psychology, 103, 708–718. Deffenbacher, J. L. (1980). Worry and emotionality in test anxiety. In I. G. Sarason (Ed.), Test anxiety: Theory, research and application (pp. 111–128). Hillside, NJ: Erlbaum. Fehr, F. S., & Stern, J. A. (1970). Peripheral physiological variables and emotion: The James–Lange theory revisited. Psychological Bulletin, 74, 411–424. Fierman, E. J., Hunt, M. F., Pratt, L. A., & Warshaw, M. G. (1993). Trauma and posttraumatic stress disorder in subjects with anxiety disorders. American Journal of Psychiatry, 150, 1872–1874. Foa, E. B., & Kozak, M. J. (1986). Emotional processing of fear: Exposure to corrective information. Psychological Bulletin, 99, 20–35. Freeston, M. H., Dugas, M. J., & Ladouceur, R. (1996). Thoughts, images, worry, and anxiety. Cognitive Therapy and Research, 20, 265–273. Freeston, M. H., Rheaume, J., Letarte, H., Dugas, M. J., & Ladouceur, R. (1994). Why do people worry? Personality and Individual Differences, 17, 791–802. Garry, M., Manning, C. G., & Loftus, E. F. (1996). Imagination inf lation: Imagining a childhood event inf lates confidence that it occurred. Psychonomic Bulletin and Review, 3, 208–214. Goldberger, A. L. (1992). Applications of chaos to physiology and medicine. In J. H. Kim & J. Stringer (Eds.), Applied chaos (pp. 321–329). New York: Wiley. Goldfried, M. R. (1971). Systematic desensitization as training in self-control. Journal of Consulting and Clinical Psychology, 37, 228–234. Goldstein, A. J., & Chambless, D. L. (1978). A reanalysis of agoraphobia. Behavior Therapy, 9, 47–59. Gray, J. A. (1982). Precis of “The neurophysiology of anxiety: An enquiry into the functions of the septo-hippocampal system. ” Behavioral and Brain Sciences, 5, 469–534. Grayson, J. B., & Borkovec, T. D. (1978). The effects of expectancy and imagined response to phobic stimuli on fear reduction. Cognitive Therapy and Research, 2, 11–24. Grayson, J. B., Foa, E. B., & Steketee, G. (1982). Habituation during exposure treatment: Distraction vs. attention-focusing. Behaviour Research and Therapy, 20, 323– 328. Greenberg, L. S., & Safran, J. D. (1987). Emotion in psychotherapy. New York: Guilford Press. Greenberg, P. E., Sisitsky, T., Kessler, R. C., Finklestein, S. N., Berndt, E. R., Davidson, J. R. T., Ballenger, J. C., & Fyer, A. J. (1999). The economic burden of anxiety disorders in the 1990s. Journal of Clinical Psychiatry, 60, 427–435.

106

THEORETICAL AND EMPIRICAL APPROACHES

Grossman, P., Stemmler, G., & Meinhardt, E. (1990). Paced respiratory sinus arrythmia as an index of cadiac parasympathetic tone during varying behavioral tasks. Psychophysiology, 27, 404–416. Grossman, P., & Svebak, S. (1987). Respiratory sinus arrhythmia as an index of parasympathetic cardiac control during active coping. Psychophysiology, 24, 228–235. Hazlett-Stevens, H., & Borkovec, T. D. (2001). Effects of worry and progressive relaxation on the reduction of fear in speech phobia: An investigation of situational exposure. Behavior Therapy, 32, 503–517. Heller, W., Nitschke, J., Etienne, M., & Miller, G. (1997). Patterns of regional brain activity differentiate types of anxiety. Journal of Abnormal Psychology, 106, 376–385. Herman, J. L., Perry, J. C., & van der Kolk, B. A. (1989). Childhood trauma in borderline personality disorder. American Journal of Psychiatry, 146, 490–495. Hoehn-Saric, R., & McLeod, D. R. (1988). The peripheral sympathetic nervous system: Its role in normal and pathological anxiety. Psychiatric Clinics of North America, 11, 375–386. Hoehn-Saric, R., McLeod, D. R., & Zimmerli, W. D. (1989). Somatic manifestations in women with generalized anxiety disorder: Physiological responses to psychological stress. Archives of General Psychiatry, 46, 1113–1119. Horowitz, L. M., Rosenberg, S. E., & Bartholomew, K. (1993). Interpersonal problems, attachment styles, and outcome in brief dynamic therapy. Journal of Consulting and Clinical Psychology, 61, 549–560. Jennings, J. R. (1986). Bodily changes during attention. In M. Coles, E. Donchin, & S. W. Porges (Eds.), Psychophysiology: Systems, processes, and applications (pp. 268–290). New York: Guilford Press. Jones, T., & Davey, G. C. (1990). The effects of cued UCS rehearsal on the retention of differential “fear” conditioning: An experimental analogue of the “worry” process. Behaviour Research and Therapy, 28, 159–164. Jordan, R. K., Schlenger, W. E., Hough, R., Kulka, R. A., Weiss, D., Fairbank, J. A., & Marmar, C. R. (1991). Lifetime and current prevalence of specific psychiatric disorders among Vietnam veterans and controls. Archives of General Psychiatry, 48, 207–215. Kosslyn, S. M. (1983). Ghosts in the mind’s machine: Creating and using images in the brain. New York: Norton. Kubzansky, L. D., Kawachi, I., Spiro, A., III, Weiss, S. T., Vokonas, P. S., & Sparrow, D. (1997). Is worrying bad for your heart?: A prospective study of worry and coronary heart disease in the Normative Aging Study. Circulation, 95, 818–824. Lang, P. J. (1985). The cognitive psychophysiology of emotion: Fear and anxiety. In A. H. Tuma & J. D. Maser (Eds.), Anxiety and the anxiety disorders (pp. 131–170). Hillsdale, NJ: Erlbaum. Lichtenstein, J., & Cassidy, J. (1991, March). The Inventory of Adult Attachment: Validation of a new measure. Paper presented at the biennial meeting of the Society for Research in Child Development, Seattle, WA. Lyonfields, J. D., Borkovec, T. D., & Thayer, J. F. (1995). Vagal tone in generalized anxiety disorder and the effects of aversive imagery and worrisome thinking. Behavior Therapy, 26, 457–466. Marten, P. A., Brown, T. A., Barlow, D. H., Borkovec, T. D., Shear, M. K., & Lydiard, R. B. (1993). Evaluation of the ratings comprising the associated symptom criterion of DSM-III-R generalized anxiety disorder. Journal of Nervous and Mental Disease, 181, 676–682.

Avoidance Theory of Worry

107

Mathews, A. (1990). Why worry?: The cognitive function of anxiety. Behaviour Research and Therapy, 28, 455–468. McGuire, W. J., & McGuire, C. V. (1991). The content, structure, and operation of thought systems. In R. S. Wyer & T. K. Srull (Eds.), Advances in social cognition (pp. 1–78). Hillside, NJ: Erlbaum. Molina, S., Borkovec, T. D., Peasley, C., & Person, D. (1998). Content analysis of worrisome streams of consciousness in anxious and dysphoric participants. Cognitive Therapy and Research, 22, 109–123. Mowrer, O. H. (1947). On the dual nature of learning: A re-interpretation of “conditioning” and “problem-solving. ” Harvard Educational Review, 17, 102–148. Öst, L. (1987). Applied relaxation: Description of a coping technique and review of controlled studies. Behaviour Research and Therapy, 25, 397–409. Paivio, A. (1986). Mental representations: A dual coding approach. New York: Oxford University Press. Paivio, A., & Marschark, M. (1991). Integrative processing of concrete and abstract sentences. In A. Paivio (Ed.), Images in the mind: The evolution of a theory (pp. 134– 154). New York: Harvester Wheatsheaf. Peasley, C. E., Molina, S., & Borkovec, T. D. (1994, November). Empathy in generalized anxiety disorder. Paper presented at the annual meeting of the Association for Advancement of Behavior Therapy, San Diego, CA. Peasley-Miklus, C., & Vrana, S. R. (2000). Effect of worrisome and relaxing thinking on fearful emotional processing. Behaviour Research and Therapy, 38, 129–144. Pincus, A. L., & Borkovec, T. D. (1994, June). Interpersonal problems in generalized anxiety disorder: Preliminary clustering of patients’ interpersonal dysfunction. Paper presented at the annual meeting of the American Psychological Society, New York. Porges, S. W. (1992). Autonomic regulation and attention. In B. A. Campbell, H. Hayne, & R. Richardson (Eds.), Attention and information processing in infants and adults (pp. 201–223). Hillside, NJ: Erbaum. Pruzinsky, T., & Borkovec, T. D. (1990). Cognitive and personality characteristics of worriers. Behaviour Research and Therapy, 28, 507–512. Roemer, L., & Borkovec, T. D. (1994). Effects of suppressing thoughts about emotional material. Journal of Abnormal Psychology, 103, 467–474. Roemer, L., Molina, S., & Borkovec, T. D. (1997). An investigation of worry content among generally anxious individuals. Journal of Nervous and Mental Disease, 185, 314–319. Roemer, L., Molina, S., Litz, B. T., & Borkovec, T. D. (1997). Preliminary investigation of the role of previous exposure to potentially traumatizing events in generalized anxiety disorder. Depression and Anxiety, 4, 134–138. Roy-Byrne, P. P. (1996). Generalized anxiety and mixed anxiety–depression: Association with disability and health care utilization. Journal of Clinical Psychiatry, 57, 86–91. Schut, A., Pincus, A., Castonguay, L. G., Bedics, J., Kline, M., Long, D., & Seals, K. (1997, November). Perceptions of attachment and self-representations at best and worst states in generalized anxiety disorder. Poster session presented at the annual meeting of the Association for Advancement of Behavior Therapy, Miami Beach, FL. Segerstrom, S. C., Glover, D. A., Craske, M. G., & Fahey, J. L. (1999). Worry affects the immune response to phobic fear. Brain, Behavior, and Immunity, 13, 80–92. Smith, L. C. (1984). Semantic satiation affects category membership decision time but not lexical priming. Memory and Cognition, 12, 483–488.

108

THEORETICAL AND EMPIRICAL APPROACHES

Stampf l, T. G., & Levis, D. J. (1967). Essentials of implosive therapy: A learning-theorybased psychodynamic behavioral therapy. Journal of Abnormal Psychology, 72, 496– 530. Stöber, J. (1998). Worry, problem elaboration and suppression of imagery: The role of concreteness. Behaviour Research and Therapy, 36, 751–756. Stöber, J., Tepperwien, S., & Staak, M. (2000). Worrying leads to reduced concreteness of problem elaborations: Evidence for the avoidance theory of worry. Anxiety, Stress, and Coping, 13, 217–227. Suinn, R. M., & Richardson, F. (1971). Anxiety management training: A nonspecific behavior therapy program for anxiety control. Behavior Therapy, 2, 498–510. Thayer, J. F., Friedman, B. H., & Borkovec, T. D. (1996). Autonomic characteristics of generalized anxiety disorder and worry. Biological Psychiatry, 39, 255–266. Thayer, J. F., Friedman, B. H., Borkovec, T. D., Johnsen, B. H., & Molina, S. (2000). Phasic heart period reactions to cued threat and non-threat stimuli in generalized anxiety disorder. Psychophysiology, 37, 361–368. Torgersen, S. (1986). Childhood and family characteristics in panic and generalized anxiety disorders. American Journal of Psychiatry, 143, 630–632. Tucker, D. M., Antes, J. R., Stenslie, C. E., & Barnhardt, T. M. (1978). Anxiety and lateral cerebral function. Journal of Abnormal Psychology, 87, 380–383. Tucker, D. M., & Newman, J. P. (1981). Verbal versus imaginal cognitive strategies in the inhibition of emotional arousal. Cognitive Therapy and Research, 5, 197–202. Tyler, S., & Tucker, D. (1982). Anxiety and perceptual structure: Individual differences in neuropsychological function. Journal of Abnormal Psychology, 91, 210–220. Vasey, M. W., & Borkovec, T. D. (1992). A catastrophizing assessment of worrisome thoughts. Cognitive Therapy and Research, 16, 1–16. Vrana, S. R., Cuthbert, B. N., & Lang, P. J. (1986). Fear imagery and text processing. Psychophysiology, 23, 247–253. Wells, A., & Papageorgiou, C. (1995). Worry and the incubation of intrusive images following stress. Behaviour Research and Therapy, 33, 579–583. Yamas, K., Hazlett-Stevens, H., & Borkovec, M. (1997, November). Alexithymia in generalized anxiety disorder. Paper presented at the annual meeting of the Association for Advancement of Behavior Therapy, Miami Beach, FL. Zuellig, A. R., Newman, M. G., Alcaine, O. A., & Behar, E. S. (1999, November). Childhood history of post-traumatic stress disorder in adults with generalized anxiety disorder, panic disorder, and non-disordered controls. Poster presented at the annual meeting of the Association for Advancement of Behavior Therapy, Toronto.

THEORETICAL AND EMPIRICAL Information-Processing Approaches APPROACHES

CHAPTER 5

Information-Processing Approaches Assessing the Selective Functioning of Attention, Interpretation, and Retrieval COLIN MACLEOD ELIZABETH RUTHERFORD

It is well established that patients with generalized anxiety disorder (GAD) report worrying with greater frequency and intensity than is the case for most individuals in the normal population (Beck, Stanley, & Zebb, 1996; Dugas et al., 1998). Worry is a mode of cognitive processing within which threatening thoughts are overrepresented (Borkovec, 1994; Borkovec & Inz, 1990; Borkovec & Lyonfields, 1993; Borkovec, Ray, & Stöber, 1998), and it tends to be accompanied by the concomitant experience of state anxiety, even when it is directly elicited by experimental instructions (Molina, Borkovec, Peasley, & Person, 1998). Accordingly, researchers have sought to identify the patterns of information processing that give rise to this type of cognitive experience, in order to address the possibility that the causal basis of GAD symptomatology might reside in the functioning of these underlying cognitive processes. Though worry represents a characteristic feature of GAD, it is also common in nonpathological anxiety. Factor-analytic studies of state anxiety symptoms in the normal population consistently have revealed a major worry factor (Endler, Parker, Bagby, & Cox, 1991; Steptoe & Kearsley, 1990). The tendency to experience elevated state anxiety varies across the 109

110

THEORETICAL AND EMPIRICAL APPROACHES

normal population as a function of trait anxiety (see Spielberger, 1972). Therefore, the tendency of patients with GAD to worry with disproportionate frequency and intensity is shared by those members of the nonclinical population who display high levels of trait anxiety (Eysenck & van Berkum, 1992). Further evidence of an association between high trait anxiety and GAD comes from the common finding that patients with GAD typically score very highly on measures of trait anxiety (Rapee, 1991) and from the observation that a heightened level of trait anxiety represents a risk factor for the development of GAD (Eysenck, 1992). Although the nature and content of worry in patients with GAD differs little from that observed in the nonclinical population (Borkovec, Shadick, & Hopkins, 1991), patients typically report having less ability to consciously control their worrying than do nonpatients who worry (Craske, Rapee, Jackel, & Barlow, 1989; Sanderson & Barlow, 1990). This clinical feature is captured in the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV; American Psychiatric Association, 1994), which explicitly requires that worry must be experienced as difficult to control before a diagnosis of GAD can be given. Thus, although patients with GAD and members of the nonclinical population with high trait anxiety share certain cognitive characteristics, they also display potentially important cognitive differences. In this chapter, we provide an overview of research designed to illuminate the types of cognitive operations most often implicated in information-processing theories of GAD. The various experimental approaches that have been employed are critically evaluated. Reference is also made where appropriate to work that has examined patterns of selective information processing in populations with high trait anxiety, primarily to highlight similarities and potential discrepancies between the information-processing characteristics of GAD and high trait anxiety. We conclude the chapter by considering the likely causal role played by observed information biases in the mediation of anxiety symptomatology.

INFORMATION-PROCESSING ACCOUNTS OF GAD Although cognitive models of anxiety vulnerability and anxiety disorders vary greatly in terms of their architecture, they share the common feature that susceptibility to anxiety and its associated disorders is attributed to the operation of selective processing biases within the cognitive system (see MacLeod, 1999). Most commonly, such accounts consider anxiety disorders such as GAD to result from the tendency to preferentially encode threatening information, to draw threatening inferences under conditions of ambiguity, and/or to preferentially retrieve threatening information from memory.

Information-Processing Approaches

111

The inf luence exerted by the early theories of Aaron Beck (1976; Beck & Emery with Greenberg, 1985; Beck, Rush, Shaw, & Emery, 1979) and G. H. Bower (1981, 1983) in stimulating and directing research on the selective processing biases associated with anxiety cannot be overemphasized, as has been recognized elsewhere (see Mathews & MacLeod, 1994). Beck and his colleagues developed a schema model (see Graesser & Nakamura, 1982) to explain the genesis and maintenance of emotional disorders. They attributed GAD to the impact of maladaptive cognitive structures termed “danger schemata,” which automatically facilitate the encoding and retrieval of threatening information, as well as the threatening interpretation of ambiguous information. In contrast, Bower proposed that anxiety is sustained and exacerbated by the automatic spread of activation from anxiety nodes within semantic memory, which serves to “prime” threat-related information, thereby rendering it disproportionately available to other cognitive processes. Although the models of Beck and Bower were developed from different theoretical perspectives, both predict that anxious individuals should display enhanced encoding and retrieval of threatening information, and should exhibit an elevated tendency to impose threatening interpretations upon ambiguous material. Furthermore, these biases should occur quite automatically, in the sense that they should not depend upon conscious mediation. Although growing diversity has accompanied their proliferation, other cognitive models of anxiety tend to share these same characteristics. For example, in response to the constraints imposed by the emerging body of experimental findings, Williams, Watts, MacLeod, and Mathews (1988, 1997) have proposed an account of anxiety vulnerability within which particular emphasis is placed upon the degree to which different types of cognitive representations undergo what Graf and Mandler (1984) have termed “integrative processing.” This automatic process involves the temporary strengthening of a mental representation’s internal structure and serves to make that representation more accessible, in that it will come to mind more readily when only some of its features are processed. According to Williams and colleagues’ model, vulnerability to anxiety is associated with a disproportionate tendency to carry out integrative processing upon threatening mental representations. The resulting enhanced accessibility of these representations causes threat-related stimuli to “pop out” of stimulus arrays in a manner that recruits selective attention and increases the ease with which threatening interpretations of ambiguity are apprehended. Williams et al. consider anxiety symptoms to result directly from this preferential automatic processing of threatening information, which they believe causally underpins the development of GAD. Many other inf luential models place similar emphasis upon the role of selective information processing in the genesis of anxious symptomatology (see MacLeod, 1991, 1999; Mogg & Bradley, 1998).

112

THEORETICAL AND EMPIRICAL APPROACHES

Öhman (1993, 1999), for example, has developed an account of anxiety that implicates an underlying bias in the preattentive processing of threatening information. Eysenck’s (1997) unified theory of anxiety places principal importance upon hypervigilance for threat information and the tendency to impose unduly threatening interpretations upon ambiguity in explaining the development and maintenance of GAD. Mogg and Bradley (1998) have put forward a cognitive–motivational theory of anxiety, which also predicts that GAD should be associated with automatic attentional orientation toward minor threat cues. Mathews and Mackintosh (1998) have advanced a model of anxiety within which processing options appraised to be highly threatening automatically will recruit disproportionate processing resources in individuals with elevated levels of anxiety vulnerability, thereby eliciting the symptoms of anxiety. However, these authors also suggest that the degree to which such automatic processing biases inf luence emotional state may depend upon whether they can be overridden by controlled cognitive operations, which they propose is more likely for nonpatients with high trait anxiety than for patients with GAD. In a refinement of Beck’s original model, Beck and Clark (1997) similarly identify a potential role for controlled processes to modify the automatic processing biases assumed to be associated with anxiety vulnerability. This refined model continues to propose that vulnerability is characterized by the rapid automatic detection of threat stimuli, but postulates that this is followed by more controlled semantic elaboration. Beck and Clark argue that this controlled elaboration stage can be employed adaptively to moderate the anxiety elicited during earlier stages of processing— for example, by engaging in “defensive” activity. However, in patients with GAD, they consider this controlled stage to involve sustained selective processing of threatening information, now with the accompanying elaboration that elicits the conscious experience of worry. Thus, despite their differences, these various cognitive models of anxiety—and many more besides (e.g., Rapee, 2001; Wells, 1995; Woody & Rachman, 1994)—have in common the central idea that automatic biases in the selective processing of threat-related information may make an important causal contribution to the development and maintenance of GAD. Given that such a wide variety of cognitive models share the core premise that GAD symptomatology may result from selective processing biases operating automatically to favor threatening information during encoding, interpretation, and/or retrieval, it is not surprising that many researchers have now directly investigated the patterns of processing selectivity that patients with GAD display across these three classes of cognitive operation. The remainder of the chapter provides a review of the experimental literature designed to establish not only the processing biases that characterize patients with GAD, but also the degree to which these are similar to, or different from, those that characterize nonpatients

Information-Processing Approaches

113

with high trait anxiety. One possibility, highlighted within several models discussed earlier, is that GAD pathology may ref lect the inability to employ controlled processing to modify biases that, for both patients and highly trait-anxious nonpatients, originate at an automatic level of cognition. To appraise the viability of this position, we contrast, whenever possible, the patterns of processing selectivity displayed by patients with GAD and participants with high trait anxiety across tasks that preclude or encourage the use of controlled processing.

ASSESSING BIASES IN THE ENCODING OF THREATENING INFORMATION Researchers have adopted a wide array of experimental techniques to assess whether anxiety vulnerability is associated with biased patterns of selective processing during the encoding of emotional information. Often participants are directly instructed to ignore the emotional stimulus information, which is presented as a potential source of distraction while they carry out some central task. The encoding of this emotional information then is revealed by the degree to which its presence serves to disrupt, or interfere with, the execution of the principal task. Such interference approaches have provided a wealth of evidence to support the presence of an encoding advantage for threatening stimuli in patients with GAD. In one early interference study, Mathews and MacLeod (1986) presented patients with GAD and nonanxious controls with either threatening or neutral stimulus words to one ear, which these participants were instructed to ignore. Simultaneously with the presentation of this emotional information, participants were required to monitor the display of a computer, which occasionally instructed them to press a response key as quickly as possible and recorded the speed with which their response was executed. It was reasoned that performance on this central task would be most efficient to the extent that participants were able to avoid the diversion of processing resources toward the distracting auditory information. For the nonanxious controls, response latencies on the central task were equivalent regardless of the emotional valence of the auditory distractors, suggesting that these individuals could equally well avoid encoding either neutral or threatening information. However, the patients with GAD displayed a significant slowing of responses on this central task in the presence of threatening compared to neutral distractors, suggesting that this anxiety disorder is indeed characterized by a tendency to preferentially encode threatening information, even under instructions to ignore this information. This tendency for patients with GAD to demonstrate disproportionate performance disruption on a central task when trying to ignore threatening distractor information has now been confirmed repeatedly (see

114

THEORETICAL AND EMPIRICAL APPROACHES

McNally, 1995, 1998). Probably the greatest wealth of evidence has come from an interference procedure that represents an adaptation of the classic Stroop color-naming task (Stroop, 1935), which has become known as the “emotional Stroop paradigm” (see Williams, Mathews, & MacLeod, 1996). This procedure involves presenting participants with threatening and neutral words in different ink colors, and instructing them to rapidly name the colors of these words while ignoring their semantic content. The tendency to selectively encode the content of threatening stimuli should be revealed by slowed color naming. Consistent with the hypothesis that patients with GAD preferentially encode threat-related information, Mathews and MacLeod (1985) demonstrated that, compared to nonanxious individuals, such patients did indeed display disproportionately slow color-naming latencies on threat words relative to neutral words. This tendency for patients with GAD to display exaggerated colornaming interference effects on threat words has since been replicated a great many times, across studies that have employed a wide variety of stimulus presentation formats (e.g., Azaies, Granger, & Debray, 1994; Bradley, Mogg, Millar, & White, 1995; Mogg, Bradley, Williams, & Mathews, 1993; Mogg, Mathews, & Weinman, 1989). Delayed color naming of threatening words has also been observed in nonpatients with high trait anxiety (Fox, 1993; Mogg et al., 2000; Richards & Millwood, 1989). Like patients with GAD, highly trait-anxious persons also demonstrate threat interference effects across a wide variety of stimuli, spanning the full range of social and physical threat. Although the similar broad scopes of their attentional biases are consistent with the possibility that high trait anxiety and GAD may represent related conditions, it is also appropriate to note certain differences in the manifestation of these effects within each population. Unlike patients with GAD, who appear to display chronically heightened threat interference effects, individuals with high trait anxiety may not reliably display these effects unless they are exposed to a stressor. Thus, for example, we (MacLeod & Rutherford, 1992) observed no difference in the profiles of color-naming latencies shown by students high and low in trait anxiety on a computerized variant of the emotional Stroop task when these participants were tested at a time of low stress. However, as the proximity of important examinations increased, the high-anxiety students became slower to colorname threat words relative to neutral words, whereas the low-anxiety students developed the reverse tendency. This has led to the conclusion that high levels of trait anxiety are characterized not by the constant presence of a selective attentional bias toward threatening information, as is the case for patients with GAD, but rather by the tendency to develop such a biased pattern of information processing in response to stress (Mathews, 1993; Mogg & Bradley, 1998). Of course, if such a pattern of attentional selectivity does contribute to the development of GAD, then this may ex-

Information-Processing Approaches

115

plain why individuals with high trait anxiety are particularly susceptible to developing GAD, especially when exposed to extended periods of environmental stress. Although patterns of performance on interference tasks have lent considerable support to the idea that patients with GAD selectively attend to threatening information, this methodological approach inevitably has also attracted its share of criticism. One reason for concern is that such patients might show performance decrements in the presence of threatening information for reasons that have nothing to do with the selective processing of this information. For example, it has been suggested that patients and controls may afford the same processing to threatening information, but that such information may provoke a more intense anxiety response among patients (MacLeod & Mathews, 1991). Disruptions of task performance, then, may ref lect only the general cognitive impairment associated with high levels of state anxiety (see Eysenck, 1982), rather than being caused by an underlying pattern of selective processing. A second concern is that slowed color naming of threat words may result from a tendency to display attentional avoidance of threatening information, rather than attentional vigilance for such information (de Ruiter & Brusschot, 1994). For example, patients with GAD could be slow to color-name threat words because they allocate attention away from such stimuli altogether, thereby impairing the registration of other attributes such as color. The viability of such an account is reinforced by Lavy and van den Hout’s (1994) demonstration that instructing participants to attentionally avoid certain classes of words during performance of a color-naming task leads to a slowing rather than to a speeding of color naming for these stimuli. If exaggerated threat interference effects truly ref lect the enhanced encoding of threatening stimuli, then it should be possible to create circumstances under which heightened anxiety vulnerability is associated with superior task performance in the presence of threatening stimuli. All this requires is the construction of tasks within which performance would be facilitated by the enhanced encoding of threatening information. A number of such tasks have been developed, and the resulting patterns of performance provide converging support for the hypothesis that anxiety vulnerability is associated with an encoding advantage for threatening information. For example, researchers have employed the lexical decision task (Forbach, Stanners, & Hochhaus, 1974) to assess whether patients with GAD encode threat-related words with disproportionate ease. In this experimental procedure, participants are presented with letter strings and must quickly press either of two response buttons to indicate whether or not each string represents a legitimate word. The rapidity with which they can accurately classify the lexical status of genuine words ref lects the speed with which they can encode these stimuli. MacLeod and Mathews

116

THEORETICAL AND EMPIRICAL APPROACHES

(1991) found no evidence that patients with GAD, when presented with the task of classifying the lexical status of single letter strings, displayed disproportionately fast lexical decision latencies on threat words. However, when these researchers employed a double-string variant of the lexical decision task (which required participants to assign priorities to the processing of alternative stimulus options), then they did indeed find that patients with GAD, unlike nonanxious controls, displayed speeded lexical decisions on threat words. Similar findings have been reported by Mogg, Mathews, Eysenck, and May (1991). This pattern of facilitated performance suggests that GAD is associated not with an increased ability to encode threatening stimuli, but rather with a tendency to assign a disproportionately high priority to the encoding of threatening stimuli, under circumstances that present alternative processing options. More direct measures of attentional allocation have yielded further evidence that patients with GAD direct processing resources toward threatening information. For example, MacLeod, Mathews, and Tata (1986) brief ly presented patients with GAD and nonanxious controls with pairs of words, separated on the vertical axis of a computer screen. On critical trials, one member of this pair was a threatening word, and the other was a neutral word. Following the termination of each such display, a small dot probe could appear on the screen within the spatial location previously occupied by either of these two words. Participants were required to press a key whenever a probe was detected, and their response latencies were recorded. It was reasoned that if attention was drawn selectively toward the more threatening words within each stimulus pair, then probes presented to this same spatial area should be detected more rapidly than those presented to the other screen location. MacLeod and colleagues observed that patients with GAD displayed speeded detection responses for probes that appeared in the same locations as the threatening stimulus words, suggesting that GAD is associated with the selective allocation of attention toward the source of threatening information. Interestingly, the nonanxious controls tended to display the reverse pattern of detection latencies, suggesting that these individuals may have systematically diverted their attentional resources away from the locus of the threatening stimuli. Numerous variants of this attentional probe paradigm have now served to reinforce this conclusion. Most have employed verbal stimuli similar to those presented by MacLeod and colleagues (1986) and have confirmed that patients with GAD, but not nonanxious controls, reliably demonstrate speeded processing of visual probes presented in the spatial vicinity of threatening stimulus words (e.g., Mogg, Bradley, & Williams, 1995; Mogg, Mathews, & Eysenck, 1992). Other versions of the attentional probe task have employed pictorial stimuli, such as angry and neutral faces, and have demonstrated that patients with GAD also show more

Information-Processing Approaches

117

rapid responses to probes appearing in the vicinity of threatening rather than neutral pictures (Bradley, Mogg, White, Groom, & de Bono, 1999). Attentional probe studies have also yielded compelling evidence to indicate that attentional orientation toward threatening stimuli may be characteristic of elevated trait anxiety in the nonclinical population. Some of these studies have shown that participants with high trait anxiety display a relative speeding to process probes presented in the spatial vicinity of threatening verbal stimuli (Mogg, Bradley, & Hallowell, 1994), while others have demonstrated similar effects with pictorial stimuli (Bradley, Mogg, & Millar, 2000). As has been the case in emotional Stroop studies of nonpatients with high trait anxiety, evidence from some attentional probe studies suggests that this tendency to orient attention toward threatening information can remain latent in such nonpatients until evoked by stressors that elevate state anxiety (Broadbent & Broadbent, 1988; MacLeod & Mathews, 1988). For example, MacLeod and Mathews (1988) found students high and low in trait anxiety not to differ in their patterns of probe detection latencies at a time of low stress. However, as these students approached the week of an important examination, those who reported high levels of trait anxiety came to display speeded detection of probes in the vicinity of threat words, suggesting the development of attentional orientation toward such stimuli. In contrast, for students with low trait anxiety, the advent of this stressor was associated with the emergence of disproportionately slow detection latencies for probes presented in the vicinity of threat words, suggesting that these individuals came to orient attention away from such threat stimuli in response to this stressor. Once again, therefore, it would appear that high levels of trait anxiety in the nonclinical population may be associated with the latent tendency to develop the same pattern of attentional bias as the one displayed by patients with GAD, but that this tendency becomes manifest only in response to environmental stress. This style of attentional response to stress may help explain why individuals high in trait anxiety are at increased risk of developing GAD. The range of experimental methodologies that have been adopted to investigate the patterns of selective attention shown by patients with GAD continues to expand. For example, in a study cited by Mogg and Bradley (1998), Mogg, Bradley, and Millar employed eye movement measures to demonstrate that patients with GAD showed a greater probability than nonanxious controls of shifting their gaze toward images of faces that display threatening expressions. Furthermore, the patients were also found to orient their gaze toward such stimuli with disproportionate speed. Despite the great variation in the methodologies that have been utilized, the findings yielded by this diverse range of experimental paradigms converge reassuringly upon the same conclusion: Together, these findings represent very strong evidence that GAD is characterized by a dispropor-

118

THEORETICAL AND EMPIRICAL APPROACHES

tionate tendency to selectively encode a broad range of threatening information. Thus the evidence is quite consistent with the possibility that selective encoding of threatening information may indeed contribute to the development and maintenance of this clinical disorder.

ASSESSING BIASES IN THE INTERPRETATION OF AMBIGUOUS INFORMATION People’s everyday environments frequently present them with ambiguous information permitting, but not requiring, interpretations that could evoke anxiety. The off hand comment of an acquaintance, a friend’s ambivalent expression, the unexpected interest of a stranger, or newly developed somatic sensations could all be accommodated by either threatening or nonthreatening interpretations. A tendency to systematically impose more threatening interpretations on such ambiguity would give people increased grounds for worry, and would be likely to elicit elevated levels of anxiety. The possibil